Winter soil CO2 emission is a very important component of the annual carbon budgets, however, almost no information on winter CO2 emission is available from the cropland soil in northeast China. In this study, soil CO2flux was measured for a 2-year period from an ongoing tillage trial on Black soil in northeast China to quantify seasonal patterns in soil CO2 flux rate and wintertime contribution to annual soil respiration. Average soil CO2 flux rates in the winter (November to March) were between 0.64 to 1.22 g CO2 m−2 d−1, in the non-growing season (October and April) were 2.09–3.56 g CO2 m−2 d−1, whereas in the growing season (May to September) they were between 10.9 to 12.7 g CO2 m−2 d−1, with no significant differences among tillage treatments. Total winter, non-growing and growing season soil CO2 emissions were 0.28–0.45 Mg C ha−1, 0.36–0.53 Mg C ha−1, and 4.52–5.55 Mg C ha−1, respectively, among tillage treatments. The contributions of winter soil respiration to annual soil CO2emission ranged from 5.1 to 7.1%, and the non-growing season emission ranged from 11.4 to 15.2% among tillage treatments. Our results indicate that in northeast China, cropland Black soil continuously emits CO2throughout the non-growing season, and the wintertime soil respiration plays a significant role in annual soil carbon budgets. Hence winter soil CO2 emission must be taken into consideration when the role of the soil ecosystem is assessed as either a sink or source of CO2 to the atmosphere.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The soil organic carbon (SOC) pool in terrestrial soil is about three times larger than the carbon pool in the atmosphere [Lal et al., 1998], so even a small change of SOC pool can affect CO2 concentration in the atmosphere, which in turn will affect global climate change. Soil respiration is the primary pathway by which CO2 fixed by plants returns to the atmosphere [Chapin et al., 1996]. Soil respiration in cropland is the sum of heterotrophic (mainly microorganisms) and autotrophic (root) respiration [Kou et al., 2007]. Generally, there are no living plants or active roots in cropland during wintertime in northeast China due to the low temperature and frozen soil conditions. Thus, the winter soil respiration in northeast China is attributed primarily to respiration by soil microorganisms. Historically, it has been assumed that microbial activity in frozen and snow covered soils is near zero and can be neglected [Fahnestock et al., 1998]. Most soil respiration measurements are conducted during the plant growing season [Fahnestock et al., 1998] and comparatively little effort is focused on non-growing season fluxes [Hubbard et al., 2005]. It has only recently become apparent that biological activity during winter in seasonally snow covered ecosystems may exert a significant influence on biogeochemical cycling and ecosystem function [Nobrega and Grogan, 2007]. This is partly because winter snowpacks can prevent significant soil freezing allowing for continued microbial activity [Brooks et al., 1996, 1997]. Recent findings indicate that soil microbial biomass may actually peak in winter [Brooks et al., 1996; Schadt et al., 2003], suggesting the presence of a soil heterotrophic community with an unknown capability of utilizing carbon [Brooks et al., 2005]. Over half of the carbon assimilated by photosynthesis in the summer can be lost during the following winter [Hubbard et al., 2005; Monson et al., 2005] which was mainly due to the soil microbial activity in the field occurring at freezing soil temperatures [Nobrega and Grogan, 2007; Panikov and Sizova, 2007]. Sommerfeld et al. also presented evidence that the soils under alpine and sub-alpine snowpacks emit CO2 throughout the snow covered period. Microbial communities have been shown to be active under snow and this has changed the estimated global rates of biogeochemical processes beneath seasonal snowpacks [Schadt et al., 2003].
 The northeast plain is an important food production base in China, where corn (Zea mays L.) and soybean (Glycine max [L.] Merr.) are the dominant crops. Black soils (Mollisols) are the major soil type in the region accounting for 28.3% of national SOC stock with only 12.9% of the nation's agricultural land area [Pan, 1999]; accordingly, the change in SOC stock in this region could be crucial to both national soil health and the global carbon budget. However, the future of agriculture is at risk in the region because of severe soil degradation associated with conventional soil management practices. Although it is well-known that conservation tillage, particularly no-tillage, can rebuild the sustainability of soil resources, no-tillage with all crop residues remaining in soil after harvest is relatively a new practice in the region. The region is characterized by cold dry winters with varied snowfall/snow cover. These factors would likely result in different winter CO2 emissions than other regions. Since the winter is quite long (November to March) in northeast China, we hypothesize that winter soil respiration plays an important role in annual soil respiration in the region. The major objective of this study was to test this hypothesis by determining the magnitude of winter soil CO2 flux rate and its contribution to total annual soil CO2emission in a Black soil agroecosystem. Secondary objectives were to evaluate the contribution of non-growing season (October and April) soil CO2 emission to total annual soil respiration; to quantify soil temperature controlling soil respiration, and to compare soil CO2emission from soils under three tillage treatments (no-tillage, NT; moldboard plough tillage, MP; and ridge tillage, RT) and two crop rotation systems (corn-soybean in NT, RT and MP; monoculture corn in NT and MP).
2. Materials and Methods
2.1. Field Site and Cropping Systems
 This study was carried out on an ongoing tillage and crop rotation field experiment, which was initiated in fall 2001 at Dehui Experimental Station (44°12′ N, 125°33′ E) of Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, in Dehui County, Jilin Province, China. The altitude of our experiment region is about 177 m. The annual mean air temperature and annual precipitation are 4.4°C and 520 mm, respectively, over the past 30 years. More than 70% of the annual precipitation occurs in June, July, and August. The climate is a semi-humid temperate continental monsoon with a long, cold winter (November to March). The winters are dry and snow cover is sparse usually less than 0.25 m, and seasonal snow cover usually begins in November and snowmelt occurs in early April. The soil is a clay loam (Typic Hapludoll); selected physical and chemical properties are given byLiang et al. . Before the establishment of the present tillage experiment, the land had been used to grow monoculture corn under conventional tillage management for more than 10 years.
 The tillage treatments consisted of NT, MP and RT, and were arranged in a randomized complete block design with four replicates. The tillage treatment was applied to the main plot which was 10.4 m × 20 m; each main plot was split lengthwise into two 5.2 m × 20 m sub-plots. Different crop rotations were applied at the sub-plot level: corn-soybean in NT, RT and MP; monoculture corn in NT and MP; and corn-corn-soybean in NT and MP. The monoculture corn and corn-corn-soybean rotations were not applied in RT. In the corn-soybean rotation, both crops were present each year and were alternated between the two sub-plots in the same main plot. In monoculture corn, both sub-plots were planted in corn each year, and in the corn-corn-soybean rotation, both sub-plots were planted to the same crop and only one of the two crops was present each year. Three of the four replicates and five of the seven tillage and crop rotation treatments employed in the tillage and crop rotation experiment were used in the present CO2flux study. These included the corn phase of corn-soybean rotation under NT, MP and RT and monoculture corn under NT and MP; the corn-corn-soybean rotation was not included.
 Corn and soybean were planted on 2 May 2008 and 1 May 2009 at a rate of 49,000 seeds ha−1 and 280,000 seeds ha−1, respectively, in 75-cm-wide rows using a KINZE-3000 NT planter (Kinze Manufacturing, Inc., Williamsburg, IA, USA); corn and soybean were harvested on 30 September 2008 and 2 October 2009. No-tillage had no soil disturbance except for planting. Weeds were controlled using broad-spectrum herbicides. Moldboard plough tillage included one fall moldboard plough (about 20 cm in depth) after corn harvest, one disking (7.5 to 10 cm in depth) and field cultivation in spring; weeds were controlled by manual hoeing. Ridge tillage included ridging in June and smashing and returning corn stalk/roots in fall; permanent ridges were maintained at a height of approximately 10 cm, which increases the surface area by approximately 10% compared with flat tillage systems; weeds were controlled by manual hoeing. Crop residues were left on soil surface after harvest for both NT and RT treatments; residue was removed prior to fall ploughing and manually replaced after ploughing in the MP treatment. The residue was replaced in the MP treatment as part of another study to replenish soil organic matter and protect the soil from wind erosion. For corn, 100 kg N ha−1, 45.5 kg P ha−1 and 78 kg K ha−1 were applied each year as starter fertilizer and additional 50 kg N ha−1was applied as top dressing at the V-6 stage. For soybean, all fertilizers were applied as starter fertilizer and included 40 kg N ha−1, 60 kg P ha−1 and 80 kg K ha−1. Starter fertilizers for all plots were applied as sidebands concurrently with planting.
2.2. Soil CO2 Flux Rate Measurements
 Three polyvinyl chloride (PVC) collars (10 cm in diameter and 5 cm in height) were vertically inserted 3 cm into the soil surface halfway between the corn rows in each plot after planting and three days before the first measurement. The soil around the PVC outside wall was tightly compacted to prevent gas leakage. Living weeds inside the collars were carefully clipped from the soil surface. The PVC collars were left in place throughout the growing season. The collars were removed before harvest and were re-inserted into the soil surface after harvest following the same protocol as described previously; in the MP treatment, the collars were re-inserted after fall ploughing.
 Soil CO2flux rate was measured using a LI-6400 Soil CO2Flux System combined with a LI-6400-09 Soil CO2Flux Chamber (LI-COR Inc., Lincoln, NE, USA). Soil CO2 flux rate was calculated on the basis of linear increase in CO2 concentration in the chamber over time.
 Soil CO2flux rate was measured for a two-year period from 15 April 2008 to 9 April 2010; measurements were made biweekly during the growing season (May to September), and monthly during the non-growing season. The non-growing season in our study site was usually 7 months from October to April, and winter in our study site was usually 5 months from November to March when land is often covered with snow. Thus, winter is a subset of the non-growing season; we split the non-growing season into the winter (November to March) and spring-fall (October and April) two independent seasons, along with growing season, to report the results.
 Measurements were carried out between 9:00 A.M. and 11:00 A.M. which is the optimal sampling time to represent the daily average of soil CO2 flux rate according to our previous research [Shi et al., 2012]. Variations in soil temperature and moisture and their effects on soil respiration are minimal within this two-hour sampling period. We used a measurement period between 2 to 5 min for each measurement. When soil CO2 flux rates were low (i.e., <1.0 g CO2 m−2 d−1), a 5-min measurement time was used, and when the flux rates were >3.0 g CO2 m−2 d−1, only 2 min were required, which enabled us to complete the measurements of all 15 collars within the two-hour time period. Soil surface temperature change was less than 1°C during this time period.
2.3. Soil Temperature and Water Content Measurements
 Soil temperature at 5 cm and 10 cm depth near each collar was measured with bent stem thermometers at the same time as the soil respiration measurements. These thermometers are similar to regular liquid filled glass thermometers except that the stem is bent approximately 45° allowing the bulb to be placed horizontally in the soil, with the graduated section protruding from the soil for reading the temperature manually. The thermometers were installed after planting, left in place over the growing season, and removed before harvest. Volumetric soil water content at 0–22 cm depth was measured using a portable time domain reflectometry (TDR) probe (Hydrosense System, IMKO, Germany) when soil respiration was measured. Soil temperature and soil water content were not measured during the winter because the thermometers and TDR probe could not be fully inserted into the frozen soil in wintertime. Daily air temperature and precipitation from 15 April 2008 to 9 April 2010 were obtained from Dehui Meteorological Bureau and are given in Figure 1. The average growing season and non-growing season temperature is 19.1°C and 7.7°C, respectively.
2.4. Dependence of Soil Respiration on Soil Temperature
 Soil temperature and respiration data from two complete years were fitted to exponential functions given in equation (1) to describe the dependence of soil respiration on soil temperature.
where SR and t are soil CO2 flux rate and soil temperature, respectively, and α and β are regression coefficients. The temperature sensitivity (Q10) of soil respiration was calculated as:
2.5. Estimates of Annual, Winter, and Spring-Fall Season Soil CO2 Emission
 Estimates of annual, winter and spring-fall season soil CO2 emission for each tillage treatment were obtained by calculating the average CO2 flux rate between sampling dates, and computing the sum of the products of the average flux rate and time between respective sampling dates (i.e., integrating) for each measurement period [Sims and Bradford, 2001] as follows:
where Δtk= tk+1 − tk, which is the number of days between each field measurement within the season; TSR is total soil CO2 emitted in the measurement season; SRm,k is the average CO2 flux rate over the interval tk+1 − tkrecorded by the LI-6400 Soil CO2 Flux System; and n is the number of soil CO2 flux measurements made within each season.
2.6. Statistical Analysis
 The analysis of variance (ANOVA) was carried out using GLM procedure of SAS software to evaluate the significances of total CO2emitted for each of the growing, spring-fall, and winter seasons and among tillage practices. Because of the design of the two rotation studies differed, statistical analysis was conducted on each study separately. Two ANOVAs were used, including one two way with factors of season and tillage and dependent variable of season CO2flux and emission, and one one-way with factor of tillage and dependent variable of annual emission. Where main or interactive fixed effects in the model were significant, least squares mean comparisons (LSMEANS) were used to evaluate the significance of effects within a treatment. Correlation between soil CO2 flux rates and temperatures were explored by exponential regression, which was also used to calculate the temperature sensitivity (Q10). All statistical analyses were performed at a significance level of 0.05 unless otherwise indicated.
3.1. Weather Characteristics During the Field Experiment
Figure 1gives the soil temperatures at 5 cm and 10 cm depths for NT soil in corn-soybean rotation; soil temperatures at 5 cm and 10 cm under the other tillage and crop rotation treatments followed a similar trend. The mean soil temperature during the growing period was similar for all five tillage and crop rotation treatments, and was 21.5°C at 5 cm and 19.0°C at 10 cm. Growing season soil temperatures at 5 cm and 10 cm depths were relatively low in May and October and reached the maximum in mid summer, which corresponded to the change of air temperature. The variation of air temperature was much larger than soil temperature. Precipitation mostly occurred in the spring and summer, and it was relatively dry in the fall and winter in this region. Normally more than 70% of total annual precipitation occur in June, July and August (Figure 1). There was no snow cover or snow cover was less than 5 cm for almost all of our winter measurements in 2008. However, snow accumulation was around 10–20 cm in depth in winter from November 2009 to March 2010. Total precipitation as snow was 20.4 mm in the winter 2008 and 48.3 mm in the winter 2009. We did not take the soil temperature or CO2 flux rate measurements in June 2009 because of nearly continuous rain.
3.2. Magnitude of Winter and Spring-Fall Season Soil CO2 Flux Rates and Their Contributions to Annual Soil CO2 Emission
 The soil CO2 flux rate followed the distinct seasonal pattern tracking soil temperature for all tillage treatments; the flux was high during summer peaking in July whereas low flux rate occurred in winter (Figures 1 and 2). Mean daily soil CO2 flux rate varied throughout the year from 0.64 to 12.7 g CO2 m−2 d−1 among tillage treatments, and the estimated annual soil CO2 emission ranged from 5.32 to 6.37 Mg C ha−1 (Table 1). Mean winter daily soil CO2 flux rate ranged from 0.64 to 1.22 g CO2 m−2 d−1 among tillage treatments, and was 5.8–10.0% of the mean daily growing season flux rate. Total amount of winter soil CO2 emission ranged from 0.24 (MP in 2008) to 0.53 (NT in 2009) Mg C ha−1 during the two years, and the mean total winter soil CO2 emitted from the different tillage treatments during November 2008 to March 2009 (0.31 Mg CO2 ha−1, snow depth was 0–5 cm) was 19.4% lower than that from November 2009 to March 2010 (0.36 Mg C ha−1, snow depth was 10–20 cm) (data not showed in Table 2). In the spring-fall season, soil CO2 flux rate was small relative to that in growing season (Table 2 and Figure 2). Mean spring-fall season soil CO2 flux rate ranged from 2.09 to 3.56 g CO2 m−2 d−1 among tillage managements, which was 16.9–32.3% of the mean growing season soil CO2 flux rate (Table 1). Total non-growing season soil CO2 emission ranged from 0.69 to 0.81 Mg C ha−1 (Table 1). The contributions of winter (November to March) and non-growing season (October to April) soil CO2 emissions to annual soil CO2 emission were 5.1–7.0% and 11.4–15.2%, respectively, among tillage treatments (Table 1).
Table 1. Winter, Spring-Fall, and Growing Season Soil CO2 Emission (Mg CO2 ha−1) in a Mollisol Under Different Tillage Practices in Northeast China From May 2008 to April 2010a
Winter (November– March)
Spring-Fall (October and April)
Growing Season (May–September)
Two ANOVAs were used for statistical analysis, including one two-way ANOVA with factors of season and tillage and dependent variable of season CO2emission, and one one-way ANOVA with factor of tillage and dependent variable of annual emission.
3.3. Effect of Tillage and Rotation Treatments on Soil CO2 Emission
 For corn-soybean rotation, the annual soil CO2 emission was greater under NT (6.37 Mg C ha−1) and RT (6.06 Mg C ha−1) than under MP (5.44 Mg C ha−1), respectively, although the difference was not statistically significant (Table1) (P > 0.05). However, for monoculture corn, the annual soil CO2 emission was 13.5% greater under NT (6.04 Mg C ha−1) than under MP (5.32 Mg C ha−1) (Table 1) (P< 0.05). This indicates that more carbon could be emitted from the Black soil along with increasing no-till acreage in northeast China.
3.4. Dependence of Soil CO2 Flux Rate on Soil Temperature
 Soil CO2 flux rate increased exponentially with soil temperature at both of the two depths for each tillage treatment (Figure 3). Soil temperature at 5 cm and 10 cm depths explained 59.0–76.4% and 71.6–79.9% of annual temporal changes in soil CO2 flux rate among tillage treatments, respectively (Figure 3). Q10 values at 5 cm and 10 cm soil were calculated by the exponential regression between soil CO2 flux rate and soil temperature at 5 cm and 10 cm depth, respectively. Soil temperature range at 5 cm and 10 cm depths was calculated by subtracting the minimum from the maximum soil temperature at 5 cm and 10 cm respectively. Soil temperature range at 5 cm was significantly lower under NT soil than MP in both crop rotation systems (P < 0.05), while they were close to each other at 10 cm soil depth (Table 3). The Q10 values among tillage systems ranged from 2.72 to 3.47 and 3.44 to 3.87 for soil temperature at the 5 cm and 10 cm depths, respectively, and they were consistently higher for the later than the former (Table 3). Although soil temperature range at 5 cm explained 86.5% of the variation in Q10 values at 5 cm, no correlation was found between soil temperature range at 10 cm and the Q10 values at 10 cm among tillage treatments (data not shown). The Q10 values at 5 cm were significantly higher (P < 0.05) under NT than under MP, while Q10 values at 10 cm were close to each other among five tillage and crop rotation treatments (Table 3).
Table 3. Soil Organic Carbon Volumetric Soil Water Contents and Q10 Values in a Mollisol Under Different Tillage Practices in Northeast China
Soil Organic Carbon (g kg−1)
Soil Water Content 0–22 cm (vol. %)
Soil Temperature Range (°C)
Means within a column followed by the same lowercase letter in C-S or by the same uppercase letter in C-C are not significantly different according to the least square mean (LSMEANS) comparison test (P < 0.05).
4.1. The Magnitude of the Soil CO2Flux Rates During Winter and Non-growing Season
 Our measured mean winter soil CO2 flux rates (0.64 – 1.22 g CO2 m−2 d−1), except NT in corn-soybean rotation which had a higher flux rate (1.22 g CO2 m−2 d−1), were consistent with soil CO2 flux rates below 1.00 g CO2 m−2 d−1 predicted by a model for winter with mean air temperatures from −15 to −5°C in ecosystems with distinct dry seasons [Raich and Potter, 1995]. The measured mean winter soil CO2 flux rate also fell within the reported values (0.53–2.73 g CO2 m−2 d−1) in a typical alpine meadow [Sommerfeld et al., 1993]. Our data were similar to results from a study conducted in a forest-steppe ecotone in north China (0.57–0.99 g CO2 m−2 d−1 [Wang et al., 2010]) and the results (0.42–1.06 g CO2 m−2 d−1) reported by Elberling for three dominant types of vegetation. Our results were also comparable to results from other studies conducted in non-agroecosystems, including those byGilmanov et al. from the sagebrush-steppe ecosystems in Idaho and Oregon (0.68–1.31 g CO2 m−2 d−1) and Brooks et al.  in forest and meadow ecosystems (0.78–1.42 g CO2 m−2 d−1). Our estimates were generally lower than those obtained from forest ecosystems. For instance, soil CO2 flux rate during the winter averaged 1.33 g CO2 m−2 d−1in the 300-year old subalpine forests [Hubbard et al., 2005] and 1.40 ± 1.0 g CO2 m−2 d−1in a deciduous broad-leaved forest [Suzuki et al., 2006]. Much higher winter soil CO2 flux rates (e.g., 1.98, 2.43 and 2.55 g CO2 m−2 d−1) were found in other forest ecosystems [McDowell et al., 2000; Schindlbacher et al., 2007; Sommerfeld et al., 1996]. Higher near surface soil organic matter and biological activity in forest and meadow ecosystems than in cropped soils may contribute to the higher winter CO2 flux rate.
 The thickness and duration of snow cover influence the subsurface soil temperature, which may further affect winter soil CO2 flux [Elberling, 2007]. The snow cover thickness and duration varied with years in our study site and was normally between 0 and 20 cm from mid November to mid March. This small amount of snow cover could not well insulate the soil from the cold atmosphere [Schimel et al., 2006; Uchida et al., 2005], which would result in low winter soil temperature and low corresponding CO2 emission in this study region. Greater total winter soil CO2 emission in 2009 than 2008 indicates a stronger heterotrophic activity under greater snow cover in 2009 when snow depth was 10–20 cm compared with 2008 when snow depth was only 0–5 cm. Considering the greatest variations in the snow cover duration and the snow depth in northeast China [An et al., 2011], it cannot be denied that the winter soil CO2 flux reported in this study exists certain uncertainty in either space or time.
 Net non-growing season ecosystem CO2exchange is mainly attributed to soil heterotrophic respiration in the agroecosystems. A full understanding of the dynamics of soil respiration during non-growing season in agroecosystems can be helpful in estimating the carbon budget [Li et al., 2010]. Unfortunately, there is a lack of information on the non-growing season carbon exchange in agroecosystems in middle and high latitude regions. In this study, non-growing season soil CO2 flux rate, 0.03–6.63 g CO2 m−2 d−1 varied among tillage systems and had a larger variation relative to other studies (1.08–5.02 g CO2 m−2 d−1) in corn agroecosystems in Jinzhou, China (about 600 km south from our study site) [Li et al., 2010]. However, these values were far lower than the highest non-growing season soil CO2 flux rate (17.1 g CO2 m−2 d−1) in a corn-soybean rotation agroecosystem in Nebraska, USA [Verma et al., 2005]. Different climates, soil types, soil quality and farm management practices probably contribute to these differences. Soil CO2flux rate increased from low and relatively stable levels in the early to mid-winter months to a much higher level in late winter to early spring (e.g., April), and these results were similar to the findings in the forest [Frank et al., 2002; Hubbard et al., 2005; Schindlbacher et al., 2007; Wang et al., 2010], tundra [Brooks et al., 1997; Oechel et al., 1997; Zimov et al., 1996] and agriculture [Li et al., 2010; Verma et al., 2005] ecosystems. High non-growing season soil CO2 flux rates immediately prior to early spring planting time indicate that abiotic and biotic conditions are conducive to microbial activity at this time. These include higher soil temperature and greater availability of melting water [Ostroumov and Siegert, 1996], root growth [Hanson et al., 2003], increased labile carbon resulting from promotion of cell lysis in freeze-thaw events and an initial pulse (<24 h) in microbial respiration after each thaw [Schimel and Clein, 1996], the release of high-quality dissolved organic carbon from microbes killed by freezing temperature [Skogland et al., 1988], and release of winter trapped CO2 [Elberling and Brandt, 2003]. Higher soil CO2 flux rates (about 2.67 g CO2 m−2 d−1) during the non-growing season were also observed during harvest time, which was probably due to high temperature in October and the largest amount of senescent crop biomass available to microbes [Verma et al., 2005].
4.2. Contributions of Winter and Non-growing Season Soil Respiration to Annual Soil CO2 Emission
 The contribution of the winter soil respiration to annual soil CO2emission (5.1%–7.2%) in the current study was consistent with reported results in a forest-steppe ecotone in north China (3.48%–7.30% [Wang et al., 2010]) and in the subalpine forests (8% [Hubbard et al., 2005]). However, our results were lower than other studies, including studies in an Austrian mountain forest (12% [Schindlbacher et al., 2007]), and in a mixed conifer forest in Washington State (17% [McDowell et al., 2000]), and much lower than the results in Rocky Mountain National Park in Colorado (23% [Mast et al., 1998]) and in a well-drained arctic heath site in northeastern Greenland (40% [Elberling and Brandt, 2003]). The contribution of winter soil respiration to annual soil CO2 emission varies with ecosystems and may be affected by factors such as relative length of winter and growing season, temperature, snow cover, vegetation and soil properties. In spite of the variability, winter soil CO2 does contribute a significant part in annual soil CO2 emission.
 The non-growing season loss of CO2 represented 33% to 40% of the summer CO2 uptake in a northern bog [Roehm and Roulet, 2003] and 11% of the summer CO2 uptake in a corn agroecosystem [Li et al., 2010]. The high non-growing season soil CO2 emission could result from concurrent soil respiration and/or from the release of CO2 stored from the previous growing season [Roehm and Roulet, 2003]. This indicates that it is necessary to assess soil CO2emission during the non-growing season as well as during the growing season. The proportion of non-growing season soil respiration to growing season soil CO2 emission (11.4%–15.2%) in this study was lower than the reported results of 15.0%–25.0% given by Verma et al.  and 22.4% given by Li et al. in corn agroecosystems. Usually, there is no plant (crop) activity and accompanying root respiration during the non-growing season in our study area. However, global warming resulting from elevated CO2 concentration in the atmosphere causes higher winter temperatures at the middle and high latitudes of the Northern Hemisphere, which may lead to more CO2emissions via increased microbial activity from farmland during this non-growing season period.
4.3. Annual Soil CO2 Emission as Affected by Tillage Practices
 Similar to the controversy over whether or not NT increases SOC stock compared with MP, there is no consensus in soil CO2 emission between NT and MP systems. The lack of consensus could be attributed to differences in regional weather, soil types, farm management practices and the duration of measuring time. More CO2 emissions from NT soil relative to MP soil in our study was in contrast with some studies where greater CO2 emissions occurred under MP than NT soil [Al-Kaisi and Yin, 2005; Ellert and Janzen, 1999; La Scala et al., 2001; Reicosky, 1997]. However, the current results were consistent with other studies [Franzluebbers et al., 1995; Hendrix et al., 1988; Oorts et al., 2007]. No-tillage generally produces an accumulation of SOC at the soil surface due to the retention of crop residues on the surface and a concomitant lack of C input at depth [Shi et al., 2011]. Also, NT is good at maintaining soil moisture which is mainly due to the retention of crop residues on the surface and undisturbed soil [Zhang et al., 2005]. Greater SOC content at 0–5 cm depth and higher soil moisture content in the plough layer in NT compared with MP soil might contribute to the higher soil CO2 emission under NT than MP observed in this study [Bowden et al., 1998; Brooks et al., 2005; La Scala et al., 2006].
4.4. Dependence of Soil CO2 Flux Rate on Soil Temperature
 The strong correlation between soil CO2 flux rate and both air and soil temperature in the current study was consistent with other field soil respiration studies [Fernandez et al., 1993; Wang et al., 2010; Yavitt et al., 1995], and the correlation indicated that soil temperature was a good surrogate for estimating annual soil respiration [Wang et al., 2010]. Soil CO2 flux rate increased with an increase in soil temperature [Bowden et al., 1998; La Scala et al., 2006] in the range from 5 to 25°C and this effect was clearly evident in this study as well. The stronger correlation between soil CO2 flux rates and soil temperatures at 10 cm than at 5 cm may indicate that more of the soil CO2 emissions are originating at 10 cm soil depth. The soil temperature is more variable at 5 cm than 10 cm and the higher correlation between soil CO2 flux rate and soil temperature at 10 cm might also indicate that changes in soil CO2 flux rate do respond to rapid changes in soil temperature.
 Understanding the sensitivity of soil respiration to temperature change and other climatic, and soil and crop management factors makes it possible to accurately evaluate the response of terrestrial carbon balance to climatic change [Peng et al., 2009]. The Q10 is commonly used to express the relationship between soil respiration and temperature [Kirschbaum, 2006]. The Q10 values, 2.72 to 3.87 observed in this study, are well within the range (2.61–5.60) of other studies for similar temperate ecosystems [Davidson et al., 1998; Saiz et al., 2006; Wang et al., 2006]. The Q10 values at 5 cm were higher than the average Q10 values of cropland in 12 study sites in more southerly areas of China (2.13 ± 0.76 at 5 cm) [Peng et al., 2009], which probably indicates that soil respiration is more sensitive to temperature in cold, high-latitude ecosystems in northeast China [Peng et al., 2009; Rustad et al., 2001]. Higher Q10 values at 10 cm than at 5 cm indicates that soil respiration was also more sensitive to the variation of soil temperature at 10 cm than that at 5 cm depth, which is consistent with Peng et al. . Q10 values at 5 cm depth were negatively correlated with soil temperature range at 5 cm among tillage treatments, but no correlation was found between Q10 values and soil temperature range at 10 cm among tillage treatments (data not shown). A reduction in Q10 with soil temperature range at 5 cm may be attributed to direct physiological acclimation of the roots or microorganisms to the changing soil temperature [Atkin and Tjoelker, 2003], or may be a result of the different tillage treatments.
 Higher Q10 values are often found at low temperatures [Chen and Tian, 2005; Shi et al., 2006], and soil respiration tends to be more sensitive to temperature with increased soil moisture before reaching a threshold value [Wang et al., 2006]. In our current study, the mean soil temperature at 5 cm or 10 cm was similar for different tillage treatments; however, higher water content was always associated with NT soil compared with MP and RT soil. We speculate that the higher Q10 values for NT soil was due to higher soil water content in NT than MP soil. This was consistent with the Davidson et al.  and Xu and Qi  studies, which reported that soil moisture played a crucial role in temperature sensitivity of soil respiration.
 Soil CO2 emissions from agricultural soils are often quantified only in the growing season; our study measured soil CO2flux rates throughout the growing and non-growing season over a two year period. This study showed that winter (November to March) and non-growing season (October to April) soil respiration accounted for 5.1% to 7.2% and 11.4 to 15.2%, respectively, of the annual soil CO2 emission in the studied Black soil in northeast China. Near surface soil temperature change controlled seasonal variation of soil CO2 emission; when the surface soil was frozen in winter, soil CO2emissions still occur. In the Black soil arable land area in northeast China, ignoring this winter and non-growing season C loss via CO2 emission would lead to underestimates of the C loss potential. More CO2 loss was found from NT than from MP soil. Our observations and data presented here are consistent with other studies indicating that winter emission of CO2is important in the global budget of SOC. Additional site-specific field measurements are required to further refine the significance of winter soil respiration from the arable land in northeast China.
 We greatly appreciate the editor and two anonymous reviewers for their thoughtful comments that helped in improving the manuscript. This study was financially supported by Natural Science Foundation of China (40801071), National Scientific and Technical Supporting Program (2009BADB3B01), Doctoral Research Foundation in Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences (O8H2041), and the Knowledge Innovation Program of the Chinese Academy of Sciences (KZCX2-EW-QN307). We also acknowledge the PhD Research Program sponsored by Ministry of Education of the People's Republic of China and Agriculture and Agri-Food Canada.