Introduction

Temperature and moisture content are key to seasonal variations in soil respiration (Paustian et al. 2016). To facilitate relevant studies, a Q10 has been defined as the degree of temperature sensitivity of soil respiration. The assessment of Q10 has been widely used as an important parameter in biogeochemical models (Raich and Tufekcioglu 2000). However, Q10 is spatiotemporally heterogeneous, and it exhibits not only significant seasonal variations, but also those variations with latitude and vegetation types (Powlson et al. 2011). Current estimates of carbon emissions from terrestrial ecosystems and the direct impacts of climate change are thus directly influenced and/or predicted by Q10 values (Schlesinger and Andrews 2000). Furthermore, Q10 largely determines the feedback relationship between climate change and carbon cycling (Pokharel et al. 2018). Therefore, under a warming climate, it is important to understand Q10 changes and their influencing factors in order to more accurately simulate and predict the key parameters of carbon cycles and processes and to clarify the relationship between soil respiration and its temperature sensitivity.

Q10 displays marked daily, monthly, seasonal and interannual variations (Li et al. 2020). According to its influencing factors, intrinsic and apparent Q10 values are divided (Yan et al. 2019). The former is the intrinsic sensitivity of soil respiration to soil temperature without taking into account all external factors; the latter (apparent sensitivity) is the sensitivity of soil respiration to soil temperature under the natural state (Chen and Tian 2005). At present, it is difficult to integrate temperature into the intrinsic or apparent Q10 values (Zhou et al. 2015). Therefore, most of the observed values are the apparent Q10. Daily variations in soil moisture are insignificant, i.e., the growth of plant roots is almost invisible and the microbial population and community remain almost unchanged (Qin et al. 2015). However, at seasonal and interannual time-scales, Q10 values well reflect these variations, including not only the responses of soil respiration to changes in soil temperature, but also to changes in soil water content, root biomass, surficial litter input, soil microbial population, and others (Qi et al. 2014). These complex response processes increase the uncertainty in the estimation of Q10 (Schindlbacher et al. 2014). Therefore, understanding the changes in Q10 is basic for correctly estimating the feedbacks between climate change and the carbon cycle (Mahecha et al. 2010; Liu et al. 2020).

In several studies, Q10 values have varied substantially with latitude and ecosystem types (Yan et al. 2019). Comprehensive analysis indicates a high sensitivity of soil respiration rates (Rs) to temperature increases in cold environments, and the fitting between Rs and soil temperature is better than between Rs and air temperature (Yang et al. 2017). In arctic, boreal and cold temperate regions, Rs is more sensitive to temperature changes and the Q10 is higher (Yuste et al. 2010). Several measurements have indicated a Q10 of ca. 2.0 for temperate forests, while for the forested lands in northeastern China, ca. 1.3–1.8 in the growing season and 3.0–5.0 in the dormant season (Mills et al. 2011). Traditionally regarded as an important carbon sink, under a warming climate, the abovementioned regions may experience greatly elevated soil respiration and rapid and massive release of CO2 into the atmosphere, positively feeding back to climate warming (Muñoz et al. 2016). However, the estimation of Q10 is mostly based on the measurements of Rs and other parameters in a specific environment at a given time period (Chen et al. 2020). For example, the weather on days for many measurements has been mostly sunny; however, in some forests, the number of rainy days in a year predominates. Therefore, the data from continuous measurements for entire growing seasons or over many years will more accurately reflect the responses of forest soil respiration to climate changes at seasonal to decadal scales.

Soil respiration occurs in the internal environment of the soil. Current observational methods cannot directly monitor the mechanisms of soil respiration, and thus most widely used methods per se adopt the measurements of water and heat parameters to explain Rs changes (Schuur et al. 2015). However, most researchers do not have a commonly agreed depth for the measurements of soil temperature (He et al. 2017). Soil respiration occurs at different soil depths but the transmission of surficial temperature changes to soil depths has a certain time lag and exponentially dampened amplitudes, leading to greatly reduced changes in soil temperature at depth (Padarian et al. 2017). Thus, there are still three key questions remaining unanswered: (1) Are/is there a depth(s) with the highest Rs? (2) Which depth(s) yield(s) the highest Rs? and, (3) Can the correlation between Rs and soil temperature at different depths indirectly reveal this/these depth(s)?

Therefore, with these three unknowns in mind, we chose a Xing’an larch (Larix gmelinii (Rupr.) Kuzen.) forest as the research site (boreal forest ecosystem) in the northern Daxing’an Mountains of Inner Mongolia Autonomous Region, Northeast China. The Daxing’an Mountains are covered by a boreal cold temperate coniferous forest. As an important part of the forest belt in eastern Asia, it plays a key role in carbon uptake and in maintaining ecological balances. We hypothesize that the Q10 obtained by high-frequency and continuous observation may better reflect the Rs change rate over a certain period of time (including precipitation and other local or short-term hydrological events), and the spatiotemporal variation of Q10 may well reflect the influences of ground and air temperature and freeze-thaw processes on the Rs in northern coniferous forests. Based on this hypothesis, we observed the soil CO2 fluxes in situ. This study is important for understanding and predicting the soil carbon cycle and its changes in the Xing’an larch forest ecosystem in the cold temperate zone, and for regional environmental management.

Materials and methods

Study area

This study was conducted in an experimental plot of the National Field Observation and Research Station of the Daxing’an Ecosystems (NFORS-DXE; 121°30′20″–121°31′0″ E and 50°49′40″–50°51′35″ N; elevation of 791–845 m a.s.l.) in the Inner Mongolia Autonomous Region (Fig. 1), located 16 km north of Genhe City in the northern part of Northeast China. The area is characterized by a continental monsoonal climate with extensive presence of frozen ground. The multi-year average (2007–2020) mean annual air temperature was − 2.9 °C. Precipitation (60%) fell in summer (generally July and August), and the snowfall generally occurred September to the following May. The multi-year average of mean snow depth was 25 cm during 2007–2020.

Fig. 1
figure 1

Location of the study area and plot

Full size image

The experimental plot is set in a Xing’an larch forest at an elevation of about 820 m a.s.l. on a northern slope. Xing’an larch, with an average diameter of breast height (DBH) of 10 cm and an average height of 10 m dominates the boreal ecosystem. Species of Ledum palustre L. prevails in the shrub understory with an average height of 0.3 m and a cover of 39%. The forest is formed over brown coniferous forest soil, with a soil pedon layer 30–40 cm thick, a 10-cm humus layer, 1.3 ± 0.06 g·cm−3 soil bulk density, and 42.7 ± 0.9 g·kg−1 soil organic matter.

Sample settings

A 20 m × 20 m representative fixed sample plot was established, and in order to ensure the reliability and representativeness of the measurements, it was divided into sixteen 5 m × 5 m sub-plots. From these sub-plots, four sub-samples were randomly selected as the observation points for soil respiration. At these points, a 10-cm high PVC soil ring with a 20-cm diameter was pressed into the soil up to 5 cm, and the surface litter was removed. Use of the soil ring can prevent the horizontal diffusion of gas and forms a closed environment to allow for more accurate gas measurements.

Research approach and methods

A fully automatic, multi-channel, soil flux measurement system was adopted (Fig. 2), consisting of a portable greenhouse gas analyzer (UGGA) (LGR Corporation, San Jose, CA, USA) and a control unit (SF-3000) (Beijing Riga United Technology Co., Ltd., Beijing, China). This instrument performs continuous, multi-channel, high frequency measurement with excellent data continuity. At each sub-sample site, the gas inside the chamber was measured every three minutes, including a gas balance time of 30 s and a gas measurement time of 150 s. Four sub-samples were automatically measured at 12-min intervals. This selection of the measurement cycle followed that of other researchers using similar instruments (Verchot et al. 2000; Song et al. 2006; Jahn et al. 2010).

Fig. 2
figure 2

LGR automatic multi-channel soil fluxes measurement system used in this study during the period of 1 April to 30 September, 2019

Full size image

To protect the instrument from damage by extreme weather, the soil greenhouse gas observation system (SF-3000, UGGA) was placed in a steel house next to the sample plot (Fig. 2), with reliable batteries under all-weather conditions. Four fully automatic breathing chambers were placed at the soil respiration point of the instrument in order to maximize the number of observations of the natural Rs. The instrument and data were checked and maintained daily during the measurement period. The concentration of water vapor and T0cm inside the breathing chambers were automatically analyzed by the UGGA analyzer, and the Tair, T5cm, and T10cm measured simultaneously every hour by the standard weather station 20 m from the sample plot.

Data processing

The data obtained from the UGGA analyzer were soil CO2 and water vapor concentrations. In this study, the release rate of gas fluxes from the soil surface to the atmosphere is positive (+), and the absorption rate of gas fluxes from the atmosphere to the soil surface, negative (−). Changes in soil CO2 and water vapor concentrations were converted into gas fluxes using a gas fluctuation model for calculating the closed-loop fluxes (Eq. 1), and the data were screened using the six times standard deviation method (Pumpanen et al. 2004; Arias-Navarro et al. 2017). This step was carried out using the SPSS software, and the results were visualized with the Origin software. The temperature sensitivity of Rs (Q10) was calculated using Eqs. 2 and 3 (Christiansen et al. 2015).

$$ F_{C} = \frac{{10VP_{0} \left( {1 - \frac{{W_{0} }}{1000}} \right)}}{{RS\left( {T_{0} + 273.15} \right)}} \times \frac{{\partial C^{^{\prime}} }}{{\partial {\text{t}}}} $$
(1)

where, Fc is the measured gas fluxes of the soil surface in mmol·m−2·s−1; V, the total internal volume of the air chamber system in cm3; P0, the initial air pressure in the air chamber in kPa; W0, the initial water vapor concentration (WVC) of the air in the air chamber in ‰; R, the universal gas constant 8.314 Pa·m3·K−1·mol−1; S, the soil measurement area in cm2; T0, the initial air temperature in the air chamber in °C; \(\frac{{\partial C^{^{\prime}} }}{{\partial {\text{t}}}}\), the discharge rate of dry measured gas after water calibration in mmol−1·s−1;

The Q10 calculation formula:

$$ R_{s} = ae^{bT} $$
(2)
$$ Q_{10} = e^{10b} = e^{{10\left( {{\text{ln}}R_{s} - {\text{ln}}a} \right)/T}} $$
(3)

where, Rs is the soil respiration rate in μmol·m−2·s−1; T is the soil surface temperature in °C, a and b are the coefficients of the equations. Q10 values of each time scale were fitted by exponential function of CO2 fluxes and soil temperature at that time scale and then calculated with the Q10 formula.

Results

Changes in near-surface soil temperatures, water vapor concentration (WVC) and R s

The climate of the Xing’an larch forest at relatively higher latitudes (50°49′40″–50°51′35″ N) is characteristic of a cold, long winter but a moist, short summer (growing season from April to September) (Fig. 3). There were larger differences amongst Tair, T0cm, and T10cm and from May to August, the Tair was always higher than the surface temperature (T0cm). The T0cm was close to 0 °C in mid-and late-April, above 0 °C during the day and below 0 °C at night. The average daily temperature in summer increased slowly, and the highest daily average T0cm of 18.7 °C occurred on 14 July, 2019. Afterwards, the T0cm declined steadily. In September, 2019, the daily average T0cm approached 0 °C, and the nighttime surface temperature began to fall below 0 °C.

Fig. 3
figure 3

Air (Tair), soil surface (T0cm), and shallow soil temperatures at 5 and 10 cm (T5cm and T10cm, respectively) in 2019. The color bars indicate the range of measured values; The small rectangles indicate the average of the measured values; and the extended lines from the small rectangles stand for one standard deviation from average of the measured value

Full size image

Because of the thermal insulation of the soil, air temperatures and soil temperatures at the two depths displayed vertical and temporal variations. They largely peaked in July when their differences reached a minimum. In each month, variations in T10cm were always the smallest among all measured soil temperatures at T0cm, T5cm and T10cm. The result of the single factor analysis indicates no significant difference between Tair and T0cm or T5cm (P > 0.05), but a significant difference between T0cm and T5cm (P = 0.026 < 0.05) and between T0cm and T10cm (P = 0.021 < 0.05).

Monthly changes in the water vapour concentration (WVC) at the soil respiratory observation points are shown in Fig. 4. In mid-April, the WVC stabilized at 3.4‰–5.0‰, but it began to fluctuate significantly at the end of April. The WVC peaked the first time at 10.5‰ in May, which extended into June; because of increasing rain, surface WVC enriched substantially. The maximum in summer (20.8‰) occurred on 21 July 2019 and began to gradually decrease. In mid-to-late September, it was more stable (6.6‰–10.3‰).

Fig. 4
figure 4

Characteristics of water–vapor concentration, air (Tair), soil surface (T0cm) and soil temperatures (T5cm and T10cm) at 5 and 10 cm depths and soil CO2 fluxes during the period of 1 April to 30 September, 2019. a Diurnal variation in water vapor concentration; b Diurnal changes in air temperature and temperatures of soil at different depths; and c Changes in the diurnal soil CO2 fluxes

Full size image

The continuous observations of soil respiration during the period of near-surface soil thawing (April to May) revealed a very low Rs in April of less than 0.1 μmol·m−2·s−1 (Fig. 4). The Rs rapidly increased to 0.3 ± 0.1 μmol·m−2·s−1 at the end of April, and began to fluctuate sharply. At the beginning of May, it was 0.4 ± 0.2 μmol·m−2·s−1 and two short-term plumes in CO2 effluxes occurred in May. The first peak occurred May 17 and the second, May 23.

In this region, the growing season is relatively short (June to September). By calculating the daily Rs averages in the major growth period (Fig. 4), soil respiration was evidently still inactive in June. The release rate of soil CO2 was relatively stable at 1.3 ± 0.5 μmol m−2 s−1. The Rs remained unchanged at the beginning and at the end of June, but after 24 June, it began to increase significantly. In July, the upward trend of Rs was enhanced and the monthly mean Rs was 2.6 ± 0.5 μmol m−2 s−1. Soil respiration began to weaken slowly in August (monthly average Rs at 3.3 ± 1.0 μmol m−2 s−1), and declined rapidly in September (monthly average Rs at 1.9 ± 0.6 μmol m−2 s−1). The Rs was larger in July and August but lower before early June and in late September. The regional Rs also had unique monthly features. For example, the Rs was low and stable in June, while it peaked over a short period from the end of July to the beginning of August. This phenomenon may be closely related to the dynamic changes in surface temperatures (T0cm).

Correlations among R s, WVC, and near-surface soil temperatures (T 0cm, T 5cm and T 10cm)

The functional relationships between Rs and environmental variables (WVC, Tair, T0cm, T5cm, and T10cm) were established through regressions. The correlation between the WVC and soil CO2 fluxes is shown in Fig. 5. The Rs was positively correlated with the growing season WVC in 2019. The optimized function of Rs is a linear function (Rs = 0.0002WVC − 0.4953) at a significant level (R2 = 0.79, P < 0.01). There was also a significant exponential correlation between Rs and soil temperatures at different depths (Fig. 6). When T0cm started to drop below 0 °C, the Rs continued to zero. The lowest correlation was found between the Rs and Tair, with the best fitting at a significant level (Rs = 0.1880e0.1562Tair, R2 = 0.67, and P < 0.01). With increasing depth, the correlation between Rs and soil temperature gradually improved, with the highest correlation (Rs = 0.1624e0.2190T10cm, R2 = 0.83, and P < 0.01) found at a depth of 10 cm. These significant correlations indicate the important influences of near-surface soil temperatures on the temporal variability of Rs in the growing season.

Fig. 5
figure 5

Correlation analysis of soil respiration and water vapor concentration (WVC) during April–September 2019

Full size image
Fig. 6
figure 6

Correlation between soil respiration rate (Rs) and air (Tair), soil surface (T0cm) and near-surface soil (T5cm and T10cm) temperatures

Full size image

There was a close relationship between seasonal changes in Rs and dynamically changing soil temperatures, but to what degree monthly Rs is controlled by soil temperature needs further investigation. In each month, an exponential function was established for relating Rs and near-surface soil temperatures from April to September (Table 1). The correlation coefficient decreased significantly from May to August. In particular, in the rainy months of June and July, Rs and soil temperatures were poorly correlated, whereas in April and September, the correlation was better. Thus, short-term Rs changes may be affected by a variety of factors (e.g., a continuous 3-day pulse of Rs in June; an Rs decrease in late July), and the impacts from a single factor of near-surface soil temperatures only provides a limited explanation for this observation.

Table 1 Monthly average soil respiration rate (Rs) for each month and functional fitting of near-surface soil temperatures (T) at different depths (T0cm, T5cm and T10cm)
Full size table

In this study, Q10 was calculated for evaluating the sensitivity of soil respiration to near-surface soil temperatures at different depths each month using Eq. 2 (Table 2). Soil CO2 effluxes were generally sensitive to temperature changes, although the effects were slight in the early stages of near-surface soil thawing. The Q10 to T0cm was 11.5 during this period; however, that of T10cm was as high as 36.6. At the beginning of April, soil CO2 fluxes were small (< 0.1 μmol·m−2 s−1), but by the end of the month, the Rs began to gradually recover and rose to 0.3 μmol·m−2 s−1. Although CO2 fluxes in April were low, its rate of increase was significantly higher than that in other months, leading to high April Q10 values. Q10 was lower and stable in the main months of the growing season (June, July, and September). CO2 fluxes were the most sensitive to July temperatures because their increasing rate was the most rapid at relatively high, stable soil temperatures. With increasing soil depth, the amplitudes of changes in temperatures decreased exponentially, enlarging the Q10. Throughout the study period, the Q10 for soil CO2 fluxes to the T0cm, T5cm, and T10cm were 5.6, 6.3, and 7.2, respectively.

Table 2 Sensitivity (Q10) of soil respiration to soil temperatures at different depths for each month of the growing season in 2019
Full size table

In different months of the major growing season, changes in soil temperature were relatively small but the Q10 fluctuated more markedly, indicating that Q10 is determined by changes in the Rs and temperature over a specific period of time. It would also be affected by other factors, for example, the monthly average Q10 in July 2019 was significantly higher than in June and August, and it might be difficult to explain the higher Q10 values only from the perspective of soil temperature. From the WVC variation characteristics in Fig. 3, the WVC in July was highest in the growing season, indicating an important contribution of frequent rainfall and the resultant high surface WVC to the high Q10.

In order to find the daily variations in Q10, a sunny day was selected in the middle of April to September (Table 3). The daily range of soil surface temperatures on April 15 was 24.0 °C, and the ratio of maximum to minimum soil CO2 fluxes was 3.2; on May 19 it was 18.4 °C, and the ratio was 2.1. Daily Q10 were 1.9 and 1.8 in April and May, respectively, significantly higher than the monthly average (1.3) from June to September.

Table 3 Q10 values on sunny days in the middle of each month of the growing season
Full size table

Discussion

Spatiotemporal variations of Q 10

Q10 values show distinct temporal variability. At different time scales, the controlling or influencing factors for ecosystem respiration rates and Q10 may vary remarkably (Maienza et al. 2017; Wang and Wang 2017). Soil respiration is variable during the main portions of the growing season and during soil thawing. There was a significantly low variability in soil CO2 fluxes in spring and autumn and high variability in summer. Soil respiration rates (Rs) also exhibit monthly variations such as low, stable values in June and a short peak period in late July to early August.

As Q10 is controlled by different ecological processes and mechanisms, the values display complex interannual, seasonal and daily variabilities (Song et al. 2016). In this study, the daily average Q10 in the growing season of 2019 was 1.5 ± 0.2 for the larch forest soil at a relatively high latitude of about 50° N. Q10 was between 1 and 2 approximately, and the range in variation of Q10 would be very large when calculated at longer time scales, such as at a month and major growing period. For example, the Q10 at the 10-cm depth in July reached 42.0. The daily Q10 was smaller than the monthly Q10 for the same site and time because of the exclusion of the influences of climate variations, such as rain, and; this better reflected the intrinsic temperature sensitivity of soil respiration (Tarkhov et al. 2019). At a seasonal scale, Q10 values not only indicate the immediate controlling effect of temperature on soil enzyme activities, but also long-term phenological control of changes in microbial communities and root growth dynamics (Gromova et al. 2020). Because of the differences between long-term and short-term temperature sensitivities of soil respiration, attention must be given to scaling issues when calculating or using Q10 values for estimating carbon budgets.

Q10 data for total global soil respiration range from 1.3 to 3.3, with a median of 2.4, most of which are forests (Yuste et al. 2010; Aguilos et al. 2018). Q10 values are closely related to latitude (Doetterl et al. 2015), and in addition, for forest ecosystems, they are lower at low and mid-latitudes, while higher at high latitudes, such as 3.4–5.6 for the U.S. Harvard Forest ecosystem at latitude around 42° N the 4.2 of the Danish beech forest at 56° N (Janssens and Pilegaard 2003). Therefore, because of the relatively higher latitude at Genhe, Northeast China, the Q10 values of the boreal larch forest soil obtained in this study are significantly higher than that of the global average.

Effects of different factors on soil respiration and Q 10 values

In this study, higher Q10 values were found in the cold season. This is consistent with the results of the inter-monthly variations of Q10 in alternating periods of low and high temperatures from a study in north China (Wang et al. 2010). Through regression analysis of monthly soil CO2 and temperature, it was recognized that a single soil temperature factor cannot fully explain short-term changes in soil CO2 fluxes, such as the large three-day pulse in June (Fig. 3). Q10 values could also be indirectly affected by the frequency and intervals of soil respiration measurements (Chen et al. 2020). In a measurement process without a specific period (freeze thaw period, rainy season) and long measurement intervals, the calculated Q10 values tend to be lower (Peichl et al. 2014; Hu et al. 2016). At the same time, the observed Q10 values are also affected by soil depth (Li et al. 2020). In this study, the Q10 values were significantly higher than those of the surface mainly because the changes in temperature in deeper soils have a dampening effect compared with the surface soil, and can better reflect the real temperatures of the internal soil environment. Therefore, changes in soil respiration rates are affected by a variety of factors, but it remains unclear how many factors affect soil respiration in a coordinated way.

Conclusion

There are seasonal variations in soil respiration and soil temperature is the dominant factor, with rainfall-induced changes in water vapour concentration the main factor for short-term fluctuations in soil CO2 fluxes. Q10 differs with time scales and soil depth, values for the surface thawing period are significantly higher than those for the growing season or thawed period of surface soil. Furthermore, soil respiration rates in the early stage of near-surface soil thawing and in the middle of the main growing period are more sensitive to temperature changes. The difference in the impacts of Q10 by soil temperatures at various depths is manifested by the higher Q10 values of temperatures of the deep soil than those of the shallow soil. Daily Q10 values are significantly lower than monthly and seasonal ones; monthly and seasonal values better reflect the changes of soil respiration affected by phenology in the natural state. To improve the accuracy of Q10 estimates for simulating soil carbon source and sink in this area, a better understanding of temporal variation characteristics of Q10 is needed. Soil moisture is also a key influencing factor for influencing the temperature sensitivity of soil respiration. However, due to a lack of accurate rainfall and soil moisture data, it is difficult to adequately explain the impact of rainfall and soil moisture changes on Q10. For future studies, the monitoring of key hydroclimatic and environmental factors in the long-term monitoring of soil respiration rate and greenhouse gases emissions would be strengthened.