Soil respiration and its components along different successional stages in the boreal biome

SUN He-Yi, YAN Guo-Yong, XING Ya-Juan, WANG Qing-Gui, ,*

(1. Department of Agricultural Resource and Environment, Heilongjiang University, Harbin 150080, China；2. School of Life Sciences, Qufu Normal University, Qufu 273165, Shandong,China)

Abstract：Understanding the changes of soil respiration and its components with forest succession is essential for long-term carbon cycle models and its sequestration underground. To examine the change of soil respiration and its influencing mechanism with forest succession, four different successional stages of grassland (GL), Betula platyphylla forest (BF), Betula platyphylla and Larix gmelinii mixed forest (MF), Larix gmelinii forest (LF) were selected in boreal zone in the Greater Khingan Mountains of northeast China in 2019, and the trenching method was used to determine autotrophic (RA) and heterotrophic (RH) respiration. Our results showed that the soil respiration (RS) of GL, BF, MF and LF were 3.11, 5.68, 5.54 and 4.69 μmol·m-2·s-1, RH were 2.37, 4.56, 4.31 and 3.50 μmol·m-2·s-1, RA were 0.74, 1.12, 1.23 and 1.19 μmol·m-2·s-1, respectively, and Rs values in exclusion litter of GL, BF, MF and LF were 3.01，5.21，5.24 and 2.88 μmol·m-2·s-1, respectively. The contribution of RA to RS about GL, BF, MF, and LF was 23.79%，19.72%，22.20% and 25.37%, respectively. Successional stages had a significant effect on RS (P<0.05). Regression analysis found that RA was positively correlated with fine root biomass (P< 0.01), and that RH was significantly positively correlated with forest litter mass (P<0.05). Our results highlight the importance of succession stages to different components of soil respiration and its significance to the estimating forest carbon sink potential.

Key words：soil respiration; forest succession; biotic factors; abiotic factors

Soil respiration (RS), as the second-largest carbon flux in terrestrial ecosystems, approximately 68～98 Pg.yr-1of carbon is released through soil respiration each year[1], which is more than ten times the CO2emissions from fossil fuel combustion[2]. As an essential component of terrestrial ecosystems, RSaccounting for up to 80% of total ecosystem respiration[3], and play a key role in the global carbon cycle[4]. Small changes in soil respiration will cause variation in atmospheric CO2concentrations and leading to climate change[5]. Therefore, it is an integral part of understanding the soil respiration in the terrestrial ecosystem and global climate change. Forest soil respiration, mostly consisting of autotrophic respiration (RA) and heterotrophic respiration (RH)[6]. RAand RHare influenced by biotic and abiotic factors at different temporal and spatial scales. RAis mainly influenced by plant phenology and photosynthetic activities[7], leaf area, and primary productivity[8], fine root biomass and other factors[9]. RHis mainly related to soil matrix, such as litter above and underground and soil organic carbon content[10].

Litters in forest ecosystems are the main carbon cycling pathways of vegetation and soil, and surface litter decomposition[11]. Litter decomposition is one of the main components of the carbon cycle in the forest ecosystem. Changes in the quantity and quality of litter on the ground therefore have an important influence on the release of surface CO2flux[12]. Many studies have shown that exclusion litter can effectively decrease RS[13], and adding litter can increase RS[14], but Berryman et al. found that litter respiration (RL) exhibits an inhibitory effect under normal and double litter conditions[15]. Therefore, in our experiment, the effect of litter on different forest stands was used, so as to reveal the influence mechanism of litter on soil respiration and provide a reference for accurately assessing the contribution of litter to Rs and the carbon storage potential of the forest[16].

Forest succession is a fundamental ecological process that not only improves forest conditions and microclimate factors[17], but also changes the biogeochemical cycle[18]. Previous studies have documented that the net primary productivity of the ecosystem increased significantly in the early succession, remained stable or decreased slightly after increasing to a certain extent to achieve the energy balance of the ecosystem and eventually reaching the climax of the succession[19]. However, Yan et al. found that soil carbon flux gradually increased with forest succession, and more carbon would be allocated underground in the forest of southern China[20]. Besides, based on Odum’s theory, the Special RHand Special RAare increasingly used as an indicator of ecosystem development. As the forest approaching a climax stage of development, the ratio of respiration to biomass will decrease because biomass becomes more energy-efficient. Low Special RHand Special RAare important for soil carbon accumulation (more carbon will be retained for soil organic carbon) and microbial biomass because less carbon is lost through respiration[21].

Therefore, we selected four types of different succession stages in the Greater Khingan Mountains, Northeast China as the research objects, including grassland (GL),Betulaplatyphyllaforest (BF),BetulaplatyphyllaandLarixgmeliniimixed forest (MF), andLarixgmeliniiforest(LF). Based on Odum’s hypothesis and the results of previous research, hypotheses were poposed were popsed: 1) RSin boreal zone increased in the early stage of succession, then decreased slightly in later succession; 2) The Special RHand Special RAgradually increased in boreal zone with forest succession.

1 Methods

1.1 Site description

The field experiment is located in the Greater Khingan Mountains, Northeast China (51°05′～51°39′ N, 125°07′～125°50′ E) (Fig.1). The average annual temperature is -2.4 ℃, which belongs to the typical continental climate of the cold temperate zone. The lowest temperature is -48 ℃, and the maximum temperature in summer is 36 ℃. The annual sufficient accumulated temperature ranges from 1 400～1 600 ℃. Early frost started in September, late frost to mid-May, and frost-free period is 90～100 days. The annual average precipitation is about 500 mm, 80% of which occurs from July to August. The dominant species in the GL areGramineae,Compositeae,Leguminosae,Rosaceae,Campanulaceae,Phrolaceae. The dominant species of the BF isBetulaplatyphylla. The dominant species of the MF areBetulaplatyphyllaandLarixgemlinii. The dominant species of the LF isLarixgemlinii. The soil layer is relatively thin, of which 0～20 cm is sandy loam, and 20～40 cm has more gravel.

Fig.1 Red mark represents the geographic location of the study sit

1.2 Selection of plant communities

The succession of plant communities is mainly characterized by changes in species composition and community structure, which exhibits different community characteristics in different stages.BetulaplatyphyllaandLarixgmeliniihave almost the same life-form in boreal zone, butBetulaplatyphyllais stronger thanLarixgmeliniiin replacing herbs. Mosses and herbs hinder the regeneration ofLarixgmelinii, butBetulaplatyphyllacan quickly form large sapling forest due to its ability of germination and spread seeds. AfterBetulaplatyphyllaformed a forest environment,Larixgmeliniiseeds began to sprout and grow under theBetulaplatyphyllaforest, gradually forming a mixed forest. And then with the growth of the mixed forest, theLarixgmeliniigradually replaced theBetulaplatyphyllato restore theLarixgmeliniiforest. Therefore, a space-for-time substitution approach was adopted and natural plant communities of four types (GL, BF, MF, LF) were selected to represent the four stages of forest succession in boreal zone.

1.3 Experimental design

In the field experiment, the trenching method was used to distinguish RSinto RHand RA. The design of the plot was begun in May 2019. A total of three 20 m×20 m standard plots were set up for each succession stage. In each plot, three PVC collars were used as measurement of soil respiration (RS), three PVC collars were used as measurement of trench treatment (RH), three PVC collars were used as measurement of exclusion litter treatment (NL), and five litter collection baskets (length 1 m, width 1 m, placed 20～25 cm above the ground, 1 mm nylon mesh on the collection baskets) were used to collect litter aboveground. The excavation profile method was used to take the soil samples, which were brought back for processing and nutrient analysis.

1.4 Soil respiration

In each plot, PVC collars (10.4 cm in diameter and 10 cm high) were permanently inserted into the soil (3 cm deep) to avoid edge effects, and the position of PVC collars do not change during the entire measurement period. To ensure the accuracy of soil respiration, no living plants kept in the PVC collars during the measurement. The Li-8100 Automated Soil CO2Flux System (Li-CorInc., Lincoln, NE, USA) was used to measure soil respiration. To avoid soil disturbance and residual root effects, preliminary measurements of soil respiration were performed two months after trenching treatment in July 2019[22]. The soil respiration was measured from 9:00 am to 3:00 pm each time to best represent the average daily emissions. While the soil respiration value was measured, the temperature of the soil near the PVC collars was measured with a TSC-W soil temperature speed tester, and the instantaneous soil temperature at a distance of 5 cm below the soil surface layer was measured. At the same time, the instantaneous volumetric water content of the soil at a distance of 5 cm below the soil surface was measured with ECH2O.

1.5 Litter and root biomass

In July 2019, fine roots in the depth of 0～10 cm soil were taken with a 5 cm inner diameter earth drill. Six root samples were obtained from each plot. Fine root samples were separated and washed, and live roots were separated from dead roots based on their color, texture, and shape[23]. All roots were weighed to determine the root biomass after drying at 60 ℃ for 72 h. Litter was collected in 2019, all litters were dry at 60 ℃ for 72 h, then the biomass was measured.

1.6 Soil microbial biomass C and fine root biomass C

Six soil samples were collected from 0～10 cm soil layers, which were sieved with a 2 mm sieve and refrigerated at 4 ℃ to determine soil microbial biomass C (MBC), and fumigation extraction was used to determine soil microbial biomass C[24]. Fine root biomass C (FRC) was calculated by multiplying the fine root biomass of the constructive species by the C concentration of the root.

1.7 Data processing and analysis

The SPSS 24.0 was used for the statistical difference between different stages of the succession and months (including soil respiration, temperature, and moisture), when there were significant differences between stages and months, LSD (Limited Slip Differential) multiples Comparative Test was used. The one-way ANOVA and LSD multiple comparison tests were used to determine the differences in MBC, FRC, special RHand special RAin different successional stages. The correlation between autotrophic respiration and fine root biomass and the relationship between heterotrophic respiration and litter were analyzed with linear regression in GraphPad Prism 8.0.2. Excel 2019 was used to make figures and tables.

An exponential model was used to simulate the relationship between soil respiration rate and soil temperature[25].

R=aebT

(1)

In Eq (1), R represents the average soil respiration rate (μmol·m-2·s-1), T represents the soil temperature (℃), and a represents the soil respiration rate (μmol·m-2·s-1) at 0 ℃, b represents the temperature response coefficient.

The relationship between soil respiration and soil moisture was measured using a linear model[26].

R=aW+b

(2)

In Eq (2), R represents the average soil respiration rate (μmol·m-2·s-1),Wrepresents the soil moisture (%), a represents the moisture response coefficient, and b is the intercept.

The soil temperature sensitivity index was expressed byQ10, which refers to the multiple of the change in soil respiration rate when the temperature increases by 10 ℃[27].

Q10=e10b

(3)

In Eq (3), b represents a temperature response coefficient.

Fine root biomass was calculated as follows.

A=B(t/106g)/(π(5 cm/2)2(hm2/108cm2))

(4)

In Eq (4),Arepresents the fine root biomass (t·ha-1), andBrepresents the average dry fine root weight per soil core (g).

Special RHand Special RA[28].

(5)

(6)

In Eq (5) and (6), MBC refers to soil microbial biomass carbon (g·C·m-2) in trenches plots and FRC referred to fine root biomass C (g·C·m-2). The unit of Special RHis Mg CO2-C g-1MBC h-1, and the unit of Specific RAis Mg CO2-C g-1root-C h-1.

2 Results

2.1 Soil properties

Soil TC, TN, and TP content showed a decreasing pattern with succession stages (Table 1). Soil pH exhibited a narrow range from 4.24 to 4.33, and the range of soil bulk density (SDB) was 0.83 to 1.17 g·cm-3in the four succession stages (Table 1).

Table 1 Base information for the four types of vegetation in the study area

Succession stages had a significant effect on MBC (P< 0.05). MBC of BF was the highest, followed by GL, MF, LF, and MBC of BF was significantly higher than those of other succession stages(Fig.2a). In July, FRC decreased and then increased slightly with succession stages. Succession stage had a significant effect on FRC (P<0.01) (Fig.2b).

Fig.2 Soil microbial biomass C (a) and fine root biomass C (b) in four succession stages

2.2 Mean soil respiration rate in different successional stages

Soil respiration (RS) of the four succession stages had significantly different in July 2019. The average soil respiration rates of GL, BF, MF, and LF were 3.11 ± 0.06, 5.68 ± 0.25, 5.54 ± 0.29, and 5.23 ± 0.39 μmol·m-2·s-1, respectively. Compared with GL, BF, MF, and LF increased by 82.63%, 78.13%, 50.80%, respectively. Among the four succession types, BF had the highest RS, which was no significantly different from MF and LF, but it was significantly different from GL (P< 0.01) (Fig.3).

Fig.3 Soil respiration of four succession forests

2.3 Contribution rate of each component of soil respiration

Trenching treatment had an inhibitory effect on soil respiration in all four succession stages in July 2019. RHof GL, BF, MF, and LF were 2.37 ± 0.19, 4.56 ± 0.36, 4.31 ± 0.26 and 3.50 ± 0.14 μmol·m-2·s-1, respectively. Compared with GL, BF, MF and LF increased by 92.41%, 81.86%, 47.68%, respectively. RHin the BF was significantly higher than those in the GL and LF (Fig.4).

Fig.4 Heterotrophic respiration of four succession forests

RAvalues of GL, BF, MF, and LF were 0.74 ± 0.19, 1.12 ± 0.16, 1.23 ± 0.07 and 1.19 ± 0.18 μmol·m-2·s-1, respectively (Fig.5). The contribution of GL, BF, MF, and LF about RAto RSwere 23.79%、19.72%、22.20% and 25.37%, respectively (Fig.6), and there was no significant difference among four succession stages, but the contribution of GL was higher than those of BF, MF and LF.

Fig.5 Autotrophic respiration of four succession forests

Fig.6 Contribution of RA to RS of four succession stages

Rsvalues in NL (exclusion litter) of GL, BF, MF and LF were 3.01 ± 0.15, 5.21 ± 0.14, 5.24 ± 0.15 and 2.88 ± 0.18 μmol·m-2·s-1, respectively (Fig.7). The Rsvalue of NL was less than that in CK (background). RSvalue in NL of the LF was significantly lower than those of the GL, BF and MF.

Fig.7 Comparison of CK and NL

2.4 Relationship between soil respiration and soil temperature and moisture

For the soil moisture and temperature in CK and NR, only the value of GL was significantly higher than those of other types (Table 2).

The regression equation could be well described the correlation between respiration rate and soil temperature and moisture. Data analysis showed that the differences in soil temperature and moisture of all four succession stages were statistically significant, and the succession significantly affected the soil temperature and moisture of CK (background), NR (trenched treatment) and NL (exclusion litter) (P< 0.01) (Table 3).

Table 2 Mean soil temperature and moisture in the four types of vegetation

Table 3 Regression models for the relationships of soil respiration with soil temperature and moisture in the four types of vegetation

Different lowercase letters indicate significant differences among succession stages, and uppercase letters indicate significant difference among CK and NL(P<0.05).

2.5 Correlation analysis

According to linear regression analysis, the average rate of autotrophic respiration at different succession stages significantly positively correlated with fine root biomass (P<0.01, R2= 0.65) (Fig.8a), and the average rate of heterotrophic respiration and the biomass of litter had positively correlation (P<0.05, R2=0.57) (Fig.8b) are founded.

Fig.8 Correlations between autotrophic respiration and fine root biomass (a) and between RH and forest floor litter mass (b) of four succession stages

2.6 Special RH and Special RA

The Special RHand Special RAin the four succession stages in 2019 was calculated, and found that the Special RHand Special RAwere the lowest in GL, followed by BF, MF, and LF. The Special RHand Special RAshowed an upward trend with the forest succession stage, the value of GL was significantly lower than that of LF (Fig.9).

Fig.9 Specific soil heterotrophic respiration (RH) of four succession stages (a) and Specific soil autotrophic respiration (RA) of four succession stages (b)

3 Discussion

3.1 Soil respiration and it’s components of four succession stages

The pattern of soil respiration is increased and then decreased (Fig.3). The changing pattern of soil respiration in this experiment, echo the finding of previous studies[29]. And the changing pattern of RHis similar to soil respiration (Fig.4). This difference of the changing pattern of RSand RHmay be attributed to changes in the biological processes and their effect on environmental factors associated with natural forest succession, and RSand RHare closely related to NPP. The Fig.3 and Fig.4 showed that GL was significantly lower BF, MF, and LF, due to the relatively low stand density (only herbaceous plants), small plant size, and low amount of soil organic matter is limited RSand RH. With the succession to BF, the quantity and quality of litter was increase, RSand RHhas also increased significantly, and nutrients input the soil in the manner of litter, which improves soil fertility and influence soil C input. In the MF, coniferous species invade the broad-leaved community, microbial biomass and litter biomass in this forest were lower than in BF. With the further development of a Larix gmelinii overstorey (LF), microbial biomass and litter biomass continued to decrease and due to the growth habits of the species lead to lower relative moisture (Table 2), it led to the lower C inputs and the nutrient supply is reduced. In the Table 3, we found soil moisture was negative significantly with soil respiration in GL and LF, this may suppress the microbial metabolism result the soil respiration was lower than BF and MF.

GL was higher than BF and MF which may be attributable to the increase in surface biomass in July, the increase of fine root biomass also increased RA(Fig.8a). Furthermore, RAis also affected by environmental factors. Soil temperature also plays an essential role in the change of RA. The average RAand temperature have similar modes in this study. The high soil temperature further promotes RA. In addition, the contribute of RAto RSwas significantly lower than the contribute of RHto RSin this research, perhaps because the current study area is a nature reserve. The forest stands here featured minimal human interference, a thick litter layer and rich SOM, causing the relatively high contribution of RH, and due to the average annual temperature is low in the field experiment which may suppress the growth of roots, resulting the relatively low contribution of RA.

Litterfall is a major pathway for through its decomposition processes. Those processes are determined not only by the amount of litter present but also by the components of the litter. Differences in between forest types were greater in plots with an intact litter layer, suggesting that litter respiration makes an important contribution to the increase in RSwith the forest succession. So, litter here is not only a source of CO2itself but also influences soil CO2efflux by indirect effects on biological processes in the underlying soil.

3.2 Special RH and Special RA in four succession stages

In previous studies, respiratory metabolic quotient usually decreases with succession. However, in this study, the Special RHand Special RAwere increases with succession. Plant litter decomposition studies have shown that respiratory metabolic quotient is increases significantly with the type of litters that resist decomposition[31]. Considering the environmental reasons of the study area, the litter of herbaceous plants can rapidly decompose in the GL, Special RHshows a predictable downward trend before the litter is almost decomposed. With the further development of a Larix gmelinii overstorey, coniferous species invade the broad-leaved community, the leaves of coniferous species are mostly thick leather than the leaves of broad-leaved species, and the stratum corneum is more developed, contains more difficult-to-decompose substances, reduce microbial efficiency, which may also be the reason for the increase in Special RH. Besides, few studies have reported the ratio of autotrophic respiration to fine root biomass C during forest succession, and the conclusion in this study was contrary to the previous results. This may be due to the fraction of NPP allocated below-ground for plant roots decreased during forest succession from GL to LF because of a decreased in the stress of soil nutrients. Therefore, it is likely that in the later succession period in the boreal forest that soil C accumulation was decreased, and the efficiency of C was also decreased.

4 Conclusions

Soil respiration and its components in the boreal biomes of the Greater Khingan Mountains in China have changed continuously with successional stages. RAdecreased with succession stage and then increased with fine root biomass. Changes in the above-and underground litter also affected RH. The significant differences in RAand RHof GL also emphasized the influence of above-and underground biomass, soil organic matter, and other factors on soil respiration. Different patterns of change between the two components indicate the importance of allocating soil carbon flux. Litter has a positive effect on soil respiration. The four succession stages of Special RHand Special RAimply that there are ecological and physiological changes between plant-microbe interactions and forest succession. These results also emphasized the necessity to consider different responses to biotic and abiotic driver mechanisms to different age forests when predicting soil CO2efflux.