Introduction

As the second largest CO2 flux between terrestrial ecosystems and atmosphere after photosynthesis1, the soil CO2 flux (Fc, the exchange rate of CO2 from soil to atmosphere) is estimated to be 68 ± 4 Pg C.y−1 globally2. A small change in Fc can significantly alter the atmospheric CO2 concentration2,3 and potentially amplify global warming4,5,6. Biotic soil CO2 flux (Fb, representing soil biological respiration) ranges from 60 ± 6 gC.m−2.y−1 for tundra to 1260 ± 57 gC.m−2.y−1 for tropical moist forests2. Abiotic soil CO2 flux (Fa, originating from soil abiotic processes such as carbonate weathering and CO2 dissolution) is reported to be no more than 3–4 gC.m−2.yr−1 7. Thus, it is customary to assume that Fc is purely of biotic origin8,9 and equal to the soil respiration rate10,11,12. In this context, most researchers tended to neglect Fa in Fc studies (annually or on shorter time scale)7,13,14 and focused only on soil biotic processes8,15,16.

However, recent studies reported anomalous fluxes or negative soil CO2 fluxes that cannot be explained by any biotic processes of soil respiration (Rs)17,18,19,20,21. In contrast to the marginal contribution assumption, they demonstrated that Fa could significantly alter the temporal variation of Fc22,23. Fa might temporally dominate the terrestrial-atmosphere carbon exchange and contribute 19–68% to annual net ecosystem exchange (NEE) in semiarid shrubland24, and even created a sink larger than 100 gC.m−2.yr−1 in desert regions25,26,27. Namely, when conditions meet, Fa can account for a significant portion of Fc: up to 13% in calcareous Mojave Desert soils28, 40% in a Mediterranean region under dry soil conditions18, and more than 75% in the Dry Valleys of Antarctica29. However, only a few studies distinguished Fa from Fc, and accurate estimates of Fa at high frequency (hourly and daily timescales) are more limited22,27,29,30. Consequently, the contribution of Fa to Fc remains poorly understood, which has resulted in an intensive debate on its magnitude and mechanisms20,21,31,32,33 and thus induced biases in estimation of soil biotic processes29,30,34. Thus, reliable partitioning of Fa and Fb is of critical importance in quantifying their contribution to Fc. As soil biotic and abiotic processes may respond to different drivers and thus respond differently to climate change24, this partitioning is essential in understanding the feedback of the soil carbon cycle in response to climate change23,29,30.

Anomalous fluxes or abiotic CO2 fluxes were mainly reported in saline or alkaline soil of arid and semiarid land20,21,25, which may occupy 50% of the total land surfaces by the end of this century35. Due the imminent transition to a warmer and more arid climate36, soils in arid and semiarid areas are generally dry and expected to become drier within this century37,38,39. Thus, studies on the biotic and abiotic components of Fc in dry land soils are needed to better understand the feedback of the carbon cycle in response to climate change in arid areas. Due to inefficient leaching resulted from low precipitation, dry land soils are usually with some degree of salinity/alkalinity. Here, we used autoclaving sterilization to distinguish Fa from Fc in saline and alkaline soils in a saline desert of northwestern China. The objectives of this study were to (1) evaluate the relative contribution of Fa and Fb to Fc over the diurnal course, and (2) quantify the bias in conventional estimation of soil biotic processes. The basic hypothesis is that in dry land saline/alkaline soils, abiotic process contributes a significant portion in CO2 exchange between soil and atmosphere.

Materials and Methods

Site description

Our experiments were conducted at a field site near the Fukang Station of Desert Ecology, Chinese Academy of Sciences (44°17′N, 87°56′E and 475 m a.s.l.). The station is located at the northern foot of Tianshan Mountains and the southern edge of the Gurbantunggut Desert in Northwest China, where saline and alkaline land is widely distributed40. The climate is temperate continental: arid, hot and dry in summer and cold in winter. Mean annual temperature is 6.6°C, mean annual precipitation is 163 mm, and mean annual class-A pan evaporation is around 800–1000 mm41. Soils are clay-loam in texture, with high salinity/alkalinity and low organic matter. The topography in the experiment site is flat (slope < 1°), and the groundwater table used to be very high, but has declined to a depth of 6 m in recent years. The dominant shrub is Tamarix ramosissima Ledeb. (average canopy cover 17%). Other herbaceous species include Salsola nitraria Pall., Suaeda acuminate Moq. and Salicornia europaea Linn., with canopy coverage of 5–30%, depending on the precipitation in that year.

Soil sampling

Typically, arid land soil is highly spatially variable42,43. To minimize the complications resulting from high spatial variability and to attain repeatable results, we opted to use well-mixed soil samples rather than intact soil cores (repeated measurements conducted using the well-mixed soil samples were not independent, and thus a mixed effects model was used to account for the autocorrelation; see details in Data analysis and statistics section). Saline and alkaline soil (FAO/UNESCO classification: Solonchaks and Solonetz) samples (0–20 cm in depth, 12 soil cores each, i.e., a total of 24 soil cores; for each core, around 9.84 kg for alkaline soil, 6.78 kg for saline soil) were collected from a typical saline desert (around the station, bulk density 1.52 ± 0.05 g.cm−3) and an alkaline site (5 km away, bulk density 1.05 ± 0.03 g.cm−3), respectively. Both soils are loamy textured with low nitrate, as all the desert soils are. Given that both sampling sites contain few shallow-rooted shrubs or grass species, which rapidly decompose in this hot, arid climate, very few roots and organic debris were found in the soil samples. Each soil sample was air-dried and sieved (2-mm mesh size) to remove large stones, and kept indoor till sterilizing or measurements. The soil samples from both sites were homogenized respectively and then placed into 24 bottom-sealed stainless steel drums (21.1 cm outer diameter, 20.3 cm inner diameter and 22 cm height). For each soil type, 12 soil drums were randomly selected for subsequent autoclaving sterilization and control (6 drums per treatment). In addition, a quartz sand drum was used to test the thermal expansion and contraction effects of the soil gaseous parts.

Sterilization treatment

To discriminate Fa from Fc, the soils were treated by autoclaving sterilization in a pressure steam chamber. Due the size of the pressure steam chamber, we sterilized one soil drum at a time. For the sterilized soil, the tops of the drums were sealed with multilayers of filter paper and brown paper to prevent water infiltrating into the soil, and then sterilization was conducted in a medical autoclave for 24 h at 120 °C44. Then, each sterilized soil drum was placed in a UV-sterilized room to prevent microbial invasion and to allow the soil to equilibrate to the ambient temperature (room temperature) and atmospheric CO2 before the CO2 flux measurements started. To ensure valid comparison, the control drums filled with unsterilized soil were also covered with filter paper at the top and were maintained under ambient temperature conditions and atmospheric pressure.

CO2 flux measurements

After pre-equilibration, the soil drums were reburied in the saline field, with a 2-cm wall exposed above the soil surface to install the CO2 flux monitoring chamber. The soil surface in the drum was at the same height as the surrounding soil to maintain its temperature in accordance with natural soil temperature fluctuation. The CO2 flux was measured using an LI-8100 Automated Soil CO2 Flux System (LI-COR, Lincoln, Nebraska, USA) equipped with a long-term monitoring chamber (LI-8100L). Automated measurements of CO2 flux were made at 10-min intervals, and the time length of one measurement was set to 120 s for the low CO2 flux rates in the arid soil. To minimize the microbial invasion effect and to maintain the sterilized soil in sterile state, the CO2 flux measurement only lasted 1 day for all soil drums. Because we only had one LI-8100, the CO2 fluxes for all 25 soil drums were cross measured (one drum at a time) on clear days from August 17th to October 24th 2009. Soil temperature was automatically measured at a depth of 1 cm at the same 10-min intervals using thermocouples (HTT thermocouple, OMEGA Engineering, Inc., Stamford, CT, USA), which were placed in the surrounding soil close to each drum45. The raw CO2 flux data and temperature data were aggregated into hourly intervals.

Soil analysis

To determine the change in the soil properties following sterilization, soil samples were collected from the soil drums after the CO2 flux measurements were completed and analyzed for soil pH, soil electrical conductivity (EC), soil water content (SWC), soil organic carbon (SOC) and soil inorganic carbon (SIC).

Soil pH and EC were determined in a soil-water suspension (1:5 of soil: water ratio) using a potentiometer and an electric conductivity meter, respectively. SWC was determined using the conventional oven-drying and balance-weighing method. SOC was measured using the K2Cr2O7–H2SO4 Walkley-Black oxidation method46. SIC was determined using a modified pressure transducer method47.

Data analysis and statistics

Based on the potential sources of CO2, we assumed Fc is the combination of Fa and Fb:

$${F}_{c}={F}_{a}+{F}_{b}$$
(1)

Fc was determined by measuring the CO2 flux of the control soil (unsterilized soil), representing the total CO2 exchange rate from soil to atmosphere. Fa was determined by measuring the CO2 flux of sterilized soil, representing the CO2 flux resulting from soil abiotic processes. Fb was calculated as the difference between Fc and Fa.

In this study, the soils were sieved and did not contain roots; therefore, soil microbial respiration (Rm) was the main contributor to Fb. In light of previous studies48,49, the functional relationships between Rm and temperature are demonstrated in Fig. 1. When the temperature is below the optimum temperature (Topt), Rm is commonly modeled using van’t Hoff equation -a simple temperature exponential function9,10,50, or modified van’t Hoff equation – using F0 and Q10 as operand which was also equivalent to van’t Hoff equation50,51 (Fig. 1A). When the temperature is higher than the Topt, Rm decreases with further increases in temperature due to the deactivation energy and enzyme degradation49,52 (Fig. 1B).

$$\text{van}\mbox{'}t\,{\rm{Hoff}}:\,{F}_{b}={R}_{m}=\alpha {e}^{\beta T}\,(T\le {T}_{{\rm{opt}}})$$
(2)
$${\rm{Modified}}\,\text{van}\mbox{'}t\,{\rm{Hoff}}:{F}_{{\rm{b}}}={F}_{0}{{{\rm{Q}}}_{10}}^{\frac{(T-{T}_{0})}{10}}$$
(3)
$${{\rm{Q}}}_{10}={{\rm{e}}}^{{\rm{\beta }}\times 10}$$
(4)

where α and β are coefficients estimated by non-linear regression: α denotes the reference soil respiration at 0°C and β provides an estimate of the Q10 coefficient (Eq. 4), representing the degree of dependence of soil respiration on temperature; T is soil temperature (°C); and Topt is the optimum soil temperature(°C); where Q10 is temperature sensitivity, defined as the factor by which CO2 production increases with a 10°C rise in temperature; F0 is a basal respiration rate for the temperature T0.

Figure 1
figure 1

Schematic diagram of the relationship between soil microbial respiration (Rm) and soil temperature. (A) Relationship between Rm and temperature when the temperature is below the optimum temperature (Topt); (B) Relationship between Rm and temperature when the temperature exceeds Topt. The Topt is defined as the temperature at which the maximum rate of soil respiration occurs.

Full size image

Therefore, the functional relationship between Fb and soil temperature was used to validate the differentiation ability of the sterilization method. Diurnal patterns of Fc, Fa and Fb in the saline and alkaline soils were constructed by using the mean hourly CO2 flux measured in 6 drums. The contribution of Fa to Fc was quantified as the ratio between the absolute value of Fa and the sum of the absolute values of Fa and Fb determined as mean hourly values. The hourly CO2 flux and temperature data were used to establish the relationships between soil temperature and Fc and Fa, while the mean hourly data from 6 drums were used in the relationship of Fb to soil temperature.

In the current study, repeated measurements at the same location over time are not independent and thus need to be corrected for autocorrelation. This can be done using mixed effects modeling53. Therefore, a mixed effects model was used to compare the diurnal patterns of Fc, Fa and Fb. Linear and non-linear regression analyses were used to statistically quantify the relationships between CO2 flux and soil temperature. Significance level was set at the 5%. All statistical analyses were performed in SAS version 9.1 using Proc Mixed (SAS Institute Inc., 2004). The figures were drawn using the MATLAB, R2012a mapping software (The MathWorks Inc., USA.).

Results

Conservation of soil properties during sterilization

The properties of the control and sterilized saline and alkaline soils are presented in Table 1. There were no significant differences in the soil properties between the control and sterilized soils for each soil type (p > 0.05), which means that the sterilizing process did not alter the soil properties. Compared with the alkaline soil, before and after the sterilization, the saline soil had higher EC, SWC, SOC, SIC and lower pH values (Table 1, p < 0.05).

Table 1 Soil properties of the control and sterilized saline and alkaline soils.
Full size table

Diurnal patterns of F c, F a and F b

The mean hourly CO2 fluxes (and their standard errors) of the control and sterilized soils revealed stable diurnal variations of Fc and Fa (Fig. 2). Fc showed a pronounced unimodal diurnal pattern in both the saline and alkaline soils; the maximum value occurred at 10:00–11:00 h and the minimum at 2:00 h. While correlated with soil temperature, the diurnal pattern of Fc preceded that of soil temperature by 3 h. In the saline soil, the diurnal amplitude of Fc was 3.4 μmol.m−2.s−1 (−0.97 to 2.43 μmol.m−2.s−1), whereas in the alkaline soil, the diurnal amplitude of Fc was 1.58 μmol.m−2.s−1 (−0.51 to 1.07 μmol.m−2.s−1). Over the diurnal course, Fc was positive (CO2 released to the atmosphere) from 8:00–17:00 h but negative (CO2 taken up from the atmosphere) for the rest of the day in both the saline and alkaline soils. Despite the diurnal variation in temperature, the Fc measured in the quartz sand fluctuated around zero throughout the day (Fig. 2C).

Figure 2
figure 2

The diurnal patterns of soil CO2 flux (Fc), abiotic CO2 flux (Fa) and soil temperature. (A) Saline soil; (B) alkaline soil; (C) quartz sand. Soil temperature was measured at a depth of 1 cm. Each value represents the mean ± SE, n = 6, except for quartz sand (n = 1).

Full size image

Similar to Fc, the unimodal diurnal pattern of Fa varied from the minimum value at 22:00–0:00 h to a maximum at 11:00–12:00 h with smaller amplitude (Fig. 2). However, the peak value of Fa occurred 2 h earlier compared with soil temperature in both the saline and alkaline soils. The diurnal amplitude of Fa was 2.18 μmol.m−2.s−1 (−1.34 to 0.84 μmol.m−2.s−1) in the saline soil and 1.42μmol.m−2.s−1 (−0.82 to 0.60 μmol.m−2.s−1) in the alkaline soil. Fa was consistently negative over the diurnal course except from 10:00–16:00 h in the saline soil and 9:00–15:00 h in the alkaline soil. On hourly scale, the average contribution of Fa to Fc was 56% (range of 11–79%) in the saline soil and 58% (range of 13–77%) in the alkaline soil, indicating that Fa rivaled or even exceeded Fb in contributing to Fc over the diurnal course.

Fb was obtained by subtracting Fa from Fc. The calculated Fb was consistently positive over the diurnal course in both the saline and alkaline soils (Fig. 3). In the saline soil, Fb exhibited a unimodal diurnal pattern, which preceded the soil temperature by 4 h. The maximum value of Fb in the saline soil was 1.94 μmol.m−2.s−1 at 10:00 h and the minimum was 0.35 μmol.m−2.s−1 at 1:00 h. However, the diurnal pattern of Fb in the alkaline soil was significantly different from that in the saline soil. In the alkaline soil, the diurnal variation of Fb indicated a bi-modal pattern. In the alkaline soil, Fb increased in the morning with increasing soil temperature and reached the first peak at 10:00 h (0.68 μmol.m−2.s−1). However, as the soil temperature continued to increase, Fb exhibited a small dip from 11:00–16:00 h, and then a second peak appeared at 17:00 h (0.63 μmol.m−2.s−1). Fb decreased throughout the night as the soil temperature decreased and reached a minimum value of 0.14 μmol.m−2.s−1at 7:00 h. The diurnal amplitude of Fb was 1.59 μmol.m−2.s−1 in the saline soil and 0.54 μmol.m−2.s−1 in the alkaline soil.

Figure 3
figure 3

The diurnal patterns of biotic CO2 flux (Fb) and soil temperature. (A) Saline soil; (B) alkaline soil. Soil temperature was measured at a depth of 1 cm. Each value represents the mean ± SE, n = 6.

Full size image

The daily averages of Fc, Fa and Fb were compared to estimate the contribution of Fa and Fb to Fc on a daily scale (Fig. 4). The daily average of Fc was 0.10 μmol.m−2.s−1 in the saline soil, which was only slightly higher than the 0.00 μmol.m−2.s−1 value for the alkaline soil. The average daily Fa in the saline soil was −0.63 μmol.m−2.s−1, which was approximately 170% greater than the value measured in the alkaline soil (−0.37 μmol.m−2.s−1). Similarly, the average daily Fb in the saline soil was 195% greater than the value measured in the alkaline soil (0.72 and 0.37 μmol.m−2.s−1, respectively). The average daily Fa and Fb values were of a similar magnitude but with different signs, indicating that the CO2 released by Fb was largely offset by the CO2 taken up by Fa over a diurnal course. Moreover, due to this covering up effect of Fa, Fc was only one-seventh of Fb in saline soil, and even an order of magnitude lower than Fb in alkaline soil, which resulted in almost neutral carbon emissions over the course of a day.

Figure 4
figure 4

The daily averages of Fc, Fa and Fb in the saline and alkaline soils. Each value represents the mean ± SE, n = 6. *and **denote significant differences between the soils at α = 0.05 and α = 0.01, respectively.

Full size image

Temperature control of F c, F a and F b

The correlations between Fc and soil temperature were significant for both the saline (Fig. 5A, p < 0.01) and alkaline soils (Fig. 5B, p < 0.01), and soil temperature explained 81% and 67% of the variation in Fc in the saline and alkaline soils, respectively. Similarly, soil temperature explained 88% and 65% of the variation in Fa in the saline and alkaline soils, respectively. Furthermore, according to the fitted Fa–temperature function, there was a threshold temperature of approximately 36 °C, for both the saline and alkaline soils, where Fa changed from negative to positive with an increase in soil temperature.

Figure 5
figure 5

Relationships between soil temperature and hourly Fc and Fa. (A,C) Salinesoil; (B,D) alkaline soil. Soil temperature was measured at a depth of 1 cm.

Full size image

The relationship between hourly Fb and soil temperature in the saline soil was significantly different from that in the alkaline soil (Fig. 6). In the saline soil, Fb increased with an increase in soil temperature (Fig. 6A). An exponential function yielded the best fit for the relationship between Fb and soil temperature, explaining 54% of the variation in Fb in the saline soil over the diurnal course. However, in the alkaline soil, Fb increased with an increase in soil temperature until 28 °C, followed by a decrease with further increases in soil temperature (Fig. 6B). Thus, the effect of soil temperature on Fb in the alkaline soil was best described using a quadratic function, and soil temperature explained 45% of the diurnal variation in Fb. In general, the value of Fb was lower in the alkaline soil than in the saline soil at the same soil temperature.

Figure 6
figure 6

Relationship between soil temperature and Fb. (A) Saline soil; (B) alkaline soil. Soil temperature was measured at a depth of 1 cm.

Full size image

Discussion

The contribution of abiotic CO2 flux

The soil CO2 fluxes reported in the current study (Figs. 2 and 4) are comparable in magnitude to soil CO2 flux measured in situ in the same undisturbed saline desert region27 and in alkaline soil54, suggesting that soil sampling procedures have little effect on soil CO2 fluxes. Previous studies showed that air-drying and sieving of soil could significantly affect soil CO2 fluxes by changing the soil water content, soil structure and soil organic matter fractions55,56. The arid soil used in the current study was severely dry, low in organic matter content and loose in structure, and our results are similar to in situ soil measurements57. We obtained continuous and high frequency measurements of abiotic CO2 fluxes covering diurnal courses after autoclaving sterilization (Fig. 2) without altering soil chemical properties (Table 1), although we have to admit that steaming may change soil moisture of the sterilized soil, but in our case, this change is not significant (Table 1). In addition, soil abiotic reactions and its rate depended on the concentration of CO2 in the soil air, and after sterilization, soil biological activity was eliminated and thus stopped CO2 release into the soil. In this case, the change of soil CO2 concentration might lead to the overestimation or underestimation of the abiotic CO2 flux (shifted towards values that are more or less positive). In original soils, positive respiration means its CO2 concentration higher than atmosphere, while negative respiration means its CO2lower than atmosphere. Therefore, autoclaving sterilization is not an accurate way to distinguish abiotic CO2 flux from soil CO2 flux, but only the best way we can find after trying various method of sterilizing the soil. The results obtained can be considered as reference values not far from the truth.

Contrary to the marginal contribution over short timescales in a previous estimation7, our results demonstrated that abiotic CO2 flux accounted for more than 56% of the soil CO2 flux measured over diurnal courses in both the saline and alkaline soils (Figs. 2 and 4). The results also show that soil temperature exerted a dominant control over abiotic CO2 flux (Fig. 5), which is consistent with recent studies in Antarctica22,29. Thus, as an important component of soil CO2 flux, abiotic CO2 flux should be considered in arid ecosystem carbon budgets, which is predicted to be more susceptible to climate change36,58.

The current study also helps to clarify the controversy regarding the underlying mechanism of the abiotic CO2 flux20,21,32,33. First, photosynthesis in the autotrophic community in biological soil crusts was ruled out because sterilization eliminated soil biological activity. According to recent studies, the main abiotic interpretations involved were subterranean ventilation20,23,59, carbonate weathering17,18,19,60, and CO2 dissolution in soil water22,44,61. The zero CO2 flux measured in the quartz sand and the comparable soil CO2 flux measured in situ in the saline and alkaline soils (Fig. 2) illustrated that in the current study, subterranean ventilation resulting from thermal expansion and contraction of the soil gasphase could not be the main contributor of the abiotic CO2 flux. Moreover, the abiotic CO2 flux was positively correlated with temperature (bidirectional - CO2 was released at higher temperature and absorbed at lower temperature), and there was a difference in this process between the saline and alkaline soils (Fig. 5). Therefore, the underlying abiotic processes were temperature-regulated, reversible, physical-chemical processes that were also affected by soil salinity and alkalinity. Although both carbonate weathering and CO2 dissolution in soil water conform to these characteristics, the CO2 flux resulting from carbonate weathering is relatively small20,34,62 compared with our data. In contrast, the DIC (dissolved inorganic carbon) derived from soil CO2 dissolution is quite large and comparable to daily soil CO2 flux34,40. The modeled CO2 dissolution process produced CO2 flux values that were comparable to our results44. Thus, CO2 dissolution was the most likely mechanism underlying the abiotic CO2 flux. Although the current study used air-dried soils with a limited amount of water, but the dryness accelerated the salinity and alkalinity of the soils, which by nature, possesses a high CO2 dissolution ability44. On the other hand, we admit that further studies are needed to draw a concrete conclusion on the underlying abiotic mechanisms.

The difference between soil respiration and measured soil CO2 flux

As the soil abiotic processes produced considerable CO2 that was absorbed by/emitted from the soil during the diurnal courses (Fig. 3), direct measurements of the soil CO2 flux can be considerably different from the biological respiration rate in saline/alkaline soils. The biotic CO2 flux we obtained was positive throughout the day (Fig. 3), which confirms the nature of biological respiration (unidirectional release of CO2 to the atmosphere)8,9. Moreover, the higher biotic CO2 flux in the saline than in the alkaline soil (Fig. 4) reflected its relatively higher soil water content and soil organic carbon content (Table 1). The biotic CO2 flux–temperature relationships (Fig. 6) agreed well with previous reports49,63,64. The 28 °C optimum temperature for soil respiration was also comparable to synthesis studies conducted in seven deserts65.

Our results showed that, on daily average, the soil CO2 flux was only one-seventh of the biotic CO2 flux in saline soil, and even an order of magnitude lower in alkaline soil, due to the negative value of the abiotic CO2 flux (Fig. 4).Hence, the soil CO2 flux severely underestimated the biotic CO2 contribution. This result agreed well with a study in a semiarid soil where biological respiration was 3.8 times higher than the measured soil CO2 flux30,34 but was different to a study in Antarctic dry valley soils where the soil CO2 flux measurements overestimated the biotic CO2 flux29. In addition, the temperature response curves of the soil CO2 flux and the biotic CO2 flux were dramatically different. While the soil CO2 fluxes were linearly correlated with soil temperature (Fig. 5), the biotic CO2 flux–temperature relationships were exponential in the saline soil (Fig. 6A) and quadratic in the alkaline soil (Fig. 6B). Accordingly, using soil CO2 flux to represent the biotic respiration of arid soil would misestimate the temperature response of soil biota and mask the temperature-inhibition effect in alkaline soil, which ultimately leads to considerable bias in the responses of arid soil to climate change. The temperature-inhibition of biotic respiration in alkaline soil might come from the effect of drought51. Therefore, it is noteworthy that in other ecosystems where environmental conditions are more favorable (e.g., higher soil water content, higher soil organic content, more roots), the relative contribution of the abiotic CO2 flux would decrease with considerable increase in the soil biology activity. Moreover, the average daily soil CO2 flux in the saline soil was only slightly higher than in the alkaline soil, while the biotic CO2 flux in the saline soil was approximately 2-times higher than in the alkaline soil(Fig. 4), indicating the soil CO2 flux measurements also underestimated soil biological activity differences between the saline and alkaline soils.

In general, the soil CO2 flux measured in arid soil is not an ideal reflection of soil respiration over the diurnal course. In addition, it is important to distinguish the abiotic or biotic sources of soil CO2 flux, so that estimates of the temperature response of soil biotic CO2 flux and the dynamics of soil organic matter in arid soils are not obscured or underestimated by soil CO2 flux measurements due to co-varying soil abiotic processes.

Conclusions

The current study is an attempt to distinguish abiotic contribution from the soil CO2 flux, by autoclaving sterilization and quantifying the abiotic contributions over diurnal courses in saline and alkaline soils. The results demonstrated that the abiotic flux was an important component of the soil CO2 flux in both the saline and alkaline soils. If taken the directly measured soil CO2 flux as soil respiration, soil biological respiration might be underestimated in the saline and alkaline soils. Moreover, the dramatic difference in the temperature response between biotic and abiotic CO2 fluxes suggested that the responses of the soil CO2 flux in arid land are not simple, but a combined results of co-varied soil biotic and abiotic processes. Our study calls for a reappraisal of the understanding of the soil CO2 flux and its temperature dependence in arid or saline/alkaline land.