Seasonal variations in temperature sensitivity of soil respiration in a larch forest in the Northern Daxing’an Mountains in Northeast China

Lin Yang·Qiuliang Zhang·Zhongtao Ma·Huijun Jin·Xiaoli Chang·Sergey S.Marchenko·Valentin V.Spektor

Abstract Temperature sensitivity of respiration of forest soils is important for its responses to climate warming and for the accurate assessment of soil carbon budget.The sensitivity of temperature (T i) to soil respiration rate (R s),and Q 10 defined by e10(ln Rs?ln a)/ Ti has been used extensively for indicating the sensitivity of soil respiration.The soil respiration under a larch (Larix gmelinii) forest in the northern Daxing’an Mountains,Northeast China was observed in situ from April to September,2019 using the dynamic chamber method.Air temperatures (Tair),soil surface temperatures(T 0cm),soil temperatures at depths of 5 and 10 cm (T 5cm and T 10cm,respectively),and soil-surface water vapor concentrations were monitored at the same time.The results show a significant monthly variability in soil respiration rate in the growing season (April–September).The Q 10 at the surface and at depths of 5 and 10 cm was estimated at 5.6,6.3,and 7.2,respectively.The Q 10@10cm over the period of surface soil thawing (Q 10@10cm,thaw =36.89)were significantly higher than that of the growing season(Q 10@10cm,growth =3.82).Furthermore,the R sin the early stage of near-surface soil thawing and in the middle of the growing season is more sensitive to changes in soil temperatures.Soil temperature is thus the dominant factor for season variations in soil respiration,but rainfall is the main controller for short-term fluctuations in respiration.Thus,the higher sensitivity of soil respiration to temperature (Q 10) is found in the middle part of the growing season.The monthly and seasonal Q 10 values better reflect the responsiveness of soil respiration to changes in hydrometeorology and ground freeze-thaw processes.This study may help assess the stability of the soil carbon pool and strength of carbon fluxes in the larch forested permafrost regions in the northern Daxing’an Mountains.

Keywords Soil respiration·Dynamic chamber method·Ground thawing·Major growth period·Soil temperature

Introduction

Temperature and moisture content are key to seasonal variations in soil respiration (Paustian et al.2016).To facilitate relevant studies,aQ10has been defined as the degree of temperature sensitivity of soil respiration.The assessment ofQ10has been widely used as an important parameter in biogeochemical models (Raich and Tufekcioglu 2000).However,Q10is 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 byQ10values (Schlesinger and Andrews 2000).Furthermore,Q10largely determines the feedback relationship between climate change and carbon cycling (Pokharel et al.2018).Therefore,under a warming climate,it is important to understandQ10changes 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.

Q10displays marked daily,monthly,seasonal and interannual variations (Li et al.2020).According to its influencing factors,intrinsic and apparentQ10values 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 apparentQ10values (Zhou et al.2015).Therefore,most of the observed values are the apparentQ10.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,Q10values 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 ofQ10(Schindlbacher et al.2014).Therefore,understanding the changes inQ10is basic for correctly estimating the feedbacks between climate change and the carbon cycle (Mahecha et al.2010;Liu et al.2020).

In several studies,Q10values have varied substantially with latitude and ecosystem types (Yan et al.2019).Comprehensive analysis indicates a high sensitivity of soil respiration rates (R s) to temperature increases in cold environments,and the fitting betweenRsand soil temperature is better than betweenR sand air temperature (Yang et al.2017).In arctic,boreal and cold temperate regions,Rsis more sensitive to temperature changes and theQ10is higher(Yuste et al.2010).Several measurements have indicated aQ10of 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 CO2into the atmosphere,positively feeding back to climate warming (Mu?oz et al.2016).However,the estimation ofQ10is mostly based on the measurements ofRsand 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 methodsper seadopt the measurements of water and heat parameters to explainR schanges (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 highestR s? (2) Which depth(s) yield(s)the highestR s? and,(3) Can the correlation betweenR sand 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 theQ10obtained by highfrequency and continuous observation may better reflect theR schange rate over a certain period of time (including precipitation and other local or short-term hydrological events),and the spatiotemporal variation ofQ10may well reflect the influences of ground and air temperature and freeze-thaw processes on theR sin northern coniferous forests.Based on this hypothesis,we observed the soil CO2fluxesin 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 Location of the study area and plot

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 ofLedum palustreL.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?3soil bulk density,and 42.7 ± 0.9 g·kg?1soil 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 subsample 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 LGR automatic multichannel soil fluxes measurement system used in this study during the period of 1 April to 30 September,2019

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 naturalRs.The instrument and data were checked and maintained daily during the measurement period.The concentration of water vapor and T0cminside the breathing chambers were automatically analyzed by the UGGA analyzer,and theTair,T5cm,andT10cmmeasured 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 CO2and 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 CO2and 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 ofR s(Q10) was calculated using Eqs.2 and 3 (Christiansen et al.2015).

where,Fcis 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;W 0, 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;T 0, the initial air temperature in the air chamber in°C;,the discharge rate of dry measured gas after water calibration in mmol?1·s?1;

TheQ10calculation formula:

where,R sis the soil respiration rate in μmol·m?2·s?1;Tis the soil surface temperature in °C,aandbare the coefficients of the equations.Q10values of each time scale were fitted by exponential function of CO2fluxes and soil temperature at that time scale and then calculated with theQ10formula.

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 amongstTair,T0cm,andT10cmand from May to August,theTairwas always higher than the surface temperature (T0cm) .TheT0cmwas 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 averageT0cmof 18.7 °C occurred on 14 July,2019.Afterwards,theT0cmdeclined steadily.In September,2019,the daily averageT0cmapproached 0 °C,and the nighttime surface temperature began to fall below 0 °C.

Fig.3 Air (T air),soil surface(T 0cm),and shallow soil temperatures at 5 and 10 cm (T 5cm and T 10cm,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

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 T10cmwere always the smallest among all measured soil temperatures atT0cm,T5cmandT10cm.The result of the single factor analysis indicates no significant difference betweenTairandT0cmorT5cm(P>0.05),but a significant difference betweenT0cmandT5cm(P=0.026 <0.05) and betweenT0cmandT10cm(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 frist 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-tolate September,it was more stable (6.6‰–10.3‰).

The continuous observations of soil respiration during the period of near-surface soil thawing (April to May) revealed a verylowRsin April of lessthan 0.1μmol·m?2·s?1(Fig.4).TheRsrapidly increased to 0.3±0.1μmol·m?2·s?1at the end of April,and began to fluctuate sharply.At the beginning of May,it was 0.4 ± 0.2 μmol·m?2·s?1and two shortterm plumes in CO2effluxes 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 Rsaverages in the major growth period (Fig.4),soil respiration was evidently still inactive in June.The release rate of soil CO2was relatively stable at 1.3 ± 0.5 μmol m?2s?1.TheR sremained unchanged at the beginning and at the end of June,but after 24 June,it began to increase significantly.In July,the upward trend ofRswas enhanced and the monthly meanRswas2.6± 0.5μmol m?2s?1.Soil respiration began to weaken slowly in August(monthly averageRsat 3.3±1.0 μmol m?2s?1),and declined rapidly in September(monthly averageR sat 1.9 ± 0.6 μmol m?2s?1).TheR swas larger in July and August but lower before early June and inlate September.The regionalR salso had unique monthly features.For example,theR swas 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).

Fig.4 Characteristics of water–vapor concentration,air (Tair),soil surface (T 0cm) and soil temperatures (T 5cm 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

Correlations among Rs, WVC,and near-surface soil temperatures (T0cm,T5cm and T10cm)

The functional relationships betweenRsand environmental variables (WVC,Tair,T0cm,T5cm,andT10cm) were established through regressions.The correlation between the WVC and soil CO2fluxes is shown in Fig.5.TheRswas positively correlated with the growing season WVC in 2019.The optimized function ofRsis a linear function(R s=0.0002WVC?0.4953)at a significant level(R2=0.79,P<0.01).There was also a significant exponential correlation betweenR sand soil temperatures at different depths(Fig.6).WhenT0cmstarted to drop below 0 °C,theR scontinued to zero.The lowest correlation was found between t heRsandTair,with the best fitting at a significant level(R s=0.1880e0.1562Tair,R2=0.67,andP<0.01).With increasing depth,the correlation betweenR sand soil temperature gradually improved,with the highest correlation(R s=0.1624e0.2190T10cm,R2=0.83,andP<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 ofR sin the growing season.

Fig.5 Correlation analysis of soil respiration and water vapor concentration (WVC) during April–September 2019

Fig.6 Correlation between soil respiration rate (R s) and air (T air),soil surface (T 0cm) and near-surface soil (T 5cm and T 10cm) temperatures

There was a close relationship between seasonal changes inR sand dynamically changing soil temperatures,but to what degree monthlyR sis controlled by soil temperature needs further investigation.In each month,an exponential function was established for relatingR sand 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,R sand soil temperatures were poorly correlated,whereas in April and September,the correlation was better.Thus,short-termRschanges may be affected by a variety of factors (e.g.,a continuous 3-day pulse ofR sin June;anR sdecrease in late July),and the impacts from a single factor of near-surface soil temperatures only provides a limited explanation for this observation.

In this study,Q10was calculated for evaluating the sensitivity of soil respiration to near-surface soil temperatures at different depths each month using Eq.2 (Table 2).Soil CO2effluxes were generally sensitive to temperature changes,although the effects were slight in the early stages of near-surface soil thawing.TheQ10toT0cmwas 11.5 during this period;however,that ofT10cmwas as high as 36.6.At the beginning of April,soil CO2fluxes were small(<0.1μmol·m?2s?1),but by the end of the month,theRsbegan to gradually recover and rose to 0.3 μmol·m?2s?1.Although CO2fluxes in April were low,its rate of increase was significantly higher than that in other months,leading to high AprilQ10values.Q10was lower and stable in the main months of the growing season (June,July,and September).CO2fluxes 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 theQ10.Throughout the study period,theQ10for soil CO2fluxes to theT0cm,T5cm,andT10cmwere 5.6,6.3,and 7.2,respectively.

In different months of the major growing season,changes in soil temperature were relatively small but theQ10fluctuated more markedly,indicating thatQ10is determined by changes in theR sand temperature over a specific period of time.It would also be affected by other factors,for example,the monthly averageQ10in July 2019 was significantly higher than in June and August,and it might be difficult to explain the higherQ10values 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 highQ10.

In order to find the daily variations inQ10,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 CO2fluxes was 3.2;on May 19 it was 18.4 °C,and the ratio was 2.1.DailyQ10were 1.9 and 1.8 in April and May,respectively,significantly higher than the monthly average (1.3)from June to September.

Table 2 Sensitivity (Q 10)of soil respiration to soil temperatures at different depths for each month of the growing season in 2019

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

Discussion

Spatiotemporal variations of Q 10

Q10values show distinct temporal variability.At different time scales,the controlling or influencing factors for ecosystem respiration rates andQ10may 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 CO2fluxes in spring and autumn and high variability in summer.Soil respiration rates (R s) also exhibit monthly variations such as low,stable values in June and a short peak period in late July to early August.

AsQ10is 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 averageQ10in the growing season of 2019 was 1.5 ± 0.2 for the larch forest soil at a relatively high latitude of about 50° N.Q10was between 1 and 2 approximately,and the range in variation ofQ10would be very large when calculated at longer time scales,such as at a month and major growing period.For example,the Q10at the 10-cm depth in July reached 42.0.The dailyQ10was smaller than the monthlyQ10for 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,Q10values 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 usingQ10values for estimating carbon budgets.

Q10data 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).Q10values 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,theQ10values 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,higherQ10values were found in the cold season.This is consistent with the results of the inter-monthly variations ofQ10in alternating periods of low and high temperatures from a study in north China (Wang et al.2010).Through regression analysis of monthly soil CO2and temperature,it was recognized that a single soil temperature factor cannot fully explain short-term changes in soil CO2fluxes,such as the large three-day pulse in June (Fig.3).Q10values 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 calculatedQ10values tend to be lower (Peichl et al.2014;Hu et al.2016).At the same time,the observedQ10values are also affected by soil depth (Li et al.2020).In this study,theQ10values 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 CO2fluxes.Q10differs 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 ofQ10by soil temperatures at various depths is manifested by the higherQ10values of temperatures of the deep soil than those of the shallow soil.DailyQ10values 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 ofQ10estimates for simulating soil carbon source and sink in this area,a better understanding of temporal variation characteristics ofQ10is 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 onQ10.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.

AcknowledgementsWe thank the on-site and logistic support and meteorological data from the Daxing’an Forest Ecosystems Research Station,Genhe,Inner Mongolia,China.

Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License,which permits use,sharing,adaptation,distribution and reproduction in any medium or format,as long as you give appropriate credit to the original author(s) and the source,provide a link to the Creative Commons licence,and indicate if changes were made.The images or other third party material in this article are included in the article’s Creative Commons licence,unless indicated otherwise in a credit line to the material.If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use,you will need to obtain permission directly from the copyright holder.To view a copy of this licence,visit http:// creat iveco mmons.org/l icen ses/ by/4.0/.