Effects of environmental factors on the temporal and spatial variations in branch and leaf CO2 efflux of Larix gmelinii var. principis-rupprechtii Mayr

The CO2 efflux of branches and leaves plays an important role in ecosystem carbon balance. Using a carbon flux system, the efflux of Larix gmelinii var. principis-rupprechtii (Dahurian larch) was investigated in 27 years (immature), 31 years (near-mature), and 47 years (mature) stands at diurnal, seasonal, and spatial scales (direction and height) as well as its connection with environmental factors from May to October 2020. Diurnal variation in efflux was a single peak, and the maximum occurring between 14:00 and 16:00. Seasonal variation also exhibited a single peak, with the maximum in late July and the minimum in early October. From May to September, efflux on the south side was the largest among the three stands, and mean values on the south side of 27 year-old, 31 year-old, and 47 year-old trees were 0.50, 0.97 and 1.05 μmol·m–2·s–1, respectively. The minimum occurred on the north side. Except for the maximum in July and September in the 27 year-old stand in the middle of the canopy, the maximum efflux in the upper canopy, and the means in the 27 year-old, 31 year-old, and 47 year-old stands were 0.49, 0.96 and 1.04 μmol·m–2·s–1, respectively; the minimum occurred in the lower canopy. Temperatures and relative humidity influenced seasonal variations in efflux. Seasonal variation in temperature sensitivity coefficient (Q10) was opposite that of temperature, increasing with decreasing temperature. At the spatial scale, maximum Q10 occurred in the mid canopy. With the efflux and temperature data in different locations, it is possible to better estimate efflux variations in each stand.


Introduction
The effect of CO 2 produced by branch and leaf respiration, a component of forest respiration, on the carbon budget of forest ecosystems is significant. On a temporal scale, the CO 2 efflux of branches and leaves (E) has obvious diurnal and seasonal variations. Most current research has reported on branches and leaves separately, and the conclusions based on various tree species and habitats were different. With regards diurnal variation, some studies found that diurnal variations of leaf CO 2 efflux exhibited a single peak, and the CO 2 efflux during the day was higher than at night Wehr et al. 2016). However, there are few studies on diurnal variations in CO 2 efflux of branches. In terms of seasonal variation, there are two kinds of dynamic variation in leaf CO 2 efflux: the first is that the efflux is generally consistent with temperatures, and the peak occurs in the middle of the growing season (Bruhn et al. 2008;Catoni et al. 2013); the second is that efflux values in the early and late growing Abstract The CO 2 efflux of branches and leaves plays an important role in ecosystem carbon balance. Using a carbon flux system, the efflux of Larix gmelinii var. principisrupprechtii (Dahurian larch) was investigated in 27 years (immature), 31 years (near-mature), and 47 years (mature) stands at diurnal, seasonal, and spatial scales (direction and height) as well as its connection with environmental factors from May to October 2020. Diurnal variation in efflux was a single peak, and the maximum occurring between 14:00 and 16:00. Seasonal variation also exhibited a single peak, with the maximum in late July and the minimum in early October. From May to September, efflux on the south side was the largest among the three stands, and mean values on the south side of 27 year-old, 31 year-old, and 47 year-old trees were 0.50, 0.97 and 1.05 μmol·m -2 ·s -1 , respectively. The minimum occurred on the north side. Except for the maximum in July and September in the 27 year-old stand in the middle of the canopy, the maximum efflux in the upper seasons are higher than those in the middle of the growing season (Dungan et al. 2003). Some studies have suggested that the seasonal variation in branch CO 2 efflux is the same as the seasonal variation in temperature, reaching a maximum in July or August and then decreasing to a minimum during the dormant season (Acosta et al. 2011;Xia 2017). However, few studies have identified branch CO 2 efflux peaks in the early growing season and declines in late summer and autumn (Vose and Ryan 2010). In addition, there are some efflux variations at a spatial scale. For example, most studies have found that, along vertical gradients, E increases with canopy height Acosta et al. 2011;Weerasinghe et al. 2014). Regardless, the variation in E in different directions has not been studied.
The respiration of trees is a complex process affected by temperature, CO 2 concentrations and age (Shapiro et al. 2004;Khomik et al. 2010;Crous et al. 2017). Tree respiration is especially affected by environmental factors (Lee et al. 2005;Chi et al. 2020). Some studies have found that respiration increases with temperature within a certain range Crous et al. 2017). With increasing temperature, the increase in enzyme activity accelerates respiration; when the air temperature rises to a certain point, the respiratory substrate becomes the dominant factor affecting respiration (Atkin and Tjoelker 2003). However, respiration decreases with water deficiency (Rodríguez-Calcerrada et al. 2011;Crous et al. 2012). When water deficiency occurs, the degree of stomatal opening of the leaves changes, and intercellular CO 2 changes, thereby affecting respiration (Tcherkez et al. 2008). In addition, water deficiency also affects substrate utilization and energy demand (Atkin and Macherel 2009). Environmental factors change with time and space, and CO 2 efflux also changes. Therefore, more comprehensive studies are needed that explore the role of environmental factors in spatiotemporal variations in respiration.
The spatiotemporal variations of stem CO 2 efflux has been explored in previous studies (Zhao et al. 2018;Zhao 2020), but there are few studies on the temporal and spatial variation of CO 2 efflux of branches and leaves of trees, and the relationship between CO 2 efflux and environmental factors is not clear. To clarify these issues, we investigated the temporal and spatial variation in E for 27 year-old (immature), 31 year-old (near-mature), and 41 year-old (mature) Dahurian larch stands during the growing season (May to September) and the dormant season (October) in 2020 in north China. In this work, CO 2 efflux of branches and leaves were monitored together. The CO 2 efflux was monitored by a custom-made respiration-measuring device together with a soil carbon flux measurement system (LI-8100A: LI-COR, Lincoln, NE, USA). This study had four objectives: (1) to explore E dynamics at different time scales, including days and seasons; (2) to explore E dynamics at different spatial scales, including directions and heights; (3) to clarify the relationship between CO 2 efflux and environmental factors; and, (4) to investigate the seasonal and spatial variation in temperature sensitivity coefficient Q 10 (the change in E resulting from a 10 ℃ increase in temperature). Research on the variation of branch and leaf CO 2 efflux and the response to environmental factors will contribute to an understanding of branch and leaf CO 2 emissions, provide data and theoretical support to find ways to regulate respiration rates, reduce carbon emissions and increased carbon storage, and will serve as a reference for future research on carbon flux of forests.

Site description
This study was carried out at the Yinhe Branch of the Saihanba Mechanized Forest Farm (116°51'-117°39' E, 42°02'-42°36' N) in Weichang County, Hebei Province, China (Fig. S1). The area has a semi-humid monsoon climate. Temperature differences between day and night are large; average annual temperature is minus 1.3 °C, average annual precipitation is 460 mm, frost-free period is 65 days, and days with gale-force winds is 50. Spring and autumn are brief and winters are long and cold. The snow period can last often for seven months. There are a variety of soil types in Saihanba area, including brown, meadow, gray forest, sandy, black, and swamp soils. The horizontal distribution of soils in the area from east to west is black, gray forest, and aeolian sandy soils. The soil type of the study area is mainly thick gray forest soil with a depth that can be > 60 cm. The area under management of the Yinhe site is 19,700 ha with a forested land area of 12,000 ha. The main tree species include Larix gmelinii var. principis-rupprechtii (Mayr) Pilg., Pinus sylvestris var. mongolica Litv., Betula platyphylla Suk. L. gmelinii var. principis-rupprechtii, commonly known as the Dahurian larch, accounts for more than 90% of the cover. As the main afforestation species, Dahurian larch begins growth in mid-May and peaks in July. It drops its needles from late September to early October and enters dormancy in mid-October in the Saihanba area. The species is important for afforestation with multiple functions. It is used for timber, water conservation and landscape greening and thus has high utilization value.

Experimental design
According to the forestry industry standard of the People' s Republic of China, the age groups of Dahurian larch plantations in north China are divided as: stands from 21 to 30 years are termed middle aged forests, from 31 to 40 years are near mature forests, and from 41 to 60 years are mature forests. The experimental site included middle aged larch, near mature larch, and mature larch plantations, referred to as 27-, 31-, and 47 year-old stands. The trials were carried out in three 20 m × 20 m plots (Table 1). During the survey, the crown projection method was used to measure canopy density. The horizontal distance from the crown edge of each tree to the trunk was measured from several directions, and crown projection plotted. The ratio of the total crown projection area to the stand area was calculated from the drawings to obtain canopy density, and the calculation results were at two decimals. In view of the large number of sites to monitor, combined with the working environment (carried out on the observation platform above 9 m) and the influence of weather, one sample tree was selected as the standard from 27-, 31-, and 47 year-old plots. Average DBH, height and form factor of the standard plot were obtained, and the individual tree that conformed to these values was selected as the standard tree. Respiration meters were installed on the southern side of the upper, middle, and lower canopies of the standard tree and on the eastern, western, southern, and northern sides of the middle canopy. The species is deciduous and in the local area, leaf or needle growth begins in mid-May, reaches the vigorous growth period in July, and needles drop from late September to early October. Therefore, the monitoring of E has seasonal limitations. From May to October 2020, it was monitored from 8:00 to 18:00 every day for three days for the first ten days and the last ten days of each month. Each measurement was repeated three times, and cycling occurred every 2 h. The soil carbon flux measurement system (LI-8100A: LI-COR, Lincoln, Nebraska, US) was used which has its own measuring device to simultaneously measure temperature, relative humidity, atmospheric pressure, and air carbon dioxide concentration. At the same time, portable weather stations were set up to monitor the environmental factors in the sample plots over a long time.

CO 2 efflux measurements
The CO 2 efflux of branches and leaves were monitored together using a custom-made respiration-measuring device together with a soil carbon flux measurement system (LI-8100A). The CO 2 efflux was measured in situ, made possible by an observation platform that was constructed. The respiration-measuring device includes a respiratory ring and sealing cover. To construct the respiratory ring, a ring was cut into two open semirings, with two holes in the middle of the semiring junction to facilitate branch passage. Symmetrical holes were drilled in the upper and lower positions on both sides of the semiring so that the respiratory ring could be fixed onto a branch with nylon cable ties. Rubber sealant was used to seal the contact position between the hole and the branch. The sealing cover was made of polyvinylchloride (PVC) pipe fittings, and rubber sealant used to seal the contact part of the sealing cover and the respiratory ring to ensure airtightness. The respiratory ring was placed 1/3 the overall length of the tree branch from the trunk. Rubber sealant was used to fill the connection position between the branch and the respiratory ring, and nylon cable ties were used to tighten the respiratory ring for fixation onto the branch. Silicone adhesive was applied into the gaps and holes at the junction of the branches and respiratory rings. Any residual, dried, or exposed silicone adhesive was removed to maintain the space between the branch and the device. During measurements, the air chamber of the system was fixed onto the respiratory ring on which the sealing cover sat. The air chamber, respiratory ring and sealing cover were bound with cable ties. The seal was checked by exhaling around the chamber and the collar to ensure no rapid change in CO 2 concentration occurred in real time. After the measurements, the change in the CO 2 concentration and the coefficient of variation (CV) were analyzed to ensure the system was properly sealed. The branch was regarded as the frustum of a cone and the length of the branch in the ring and the diameter at both ends were measured. The surface area and volume of the branches were calculated according to Eq. 1, Eq. 2, and Eq. 3.The leaf area within the ring measured by an Epson scanner and ImageJ. The area and volume parameters were input into the system (LI-8100A) to facilitate accurate calculations of CO 2 efflux.
The frustum of a cone surface area equation, Eq. 1 and generatrix equation, Eq. 2 were used to estimate the surface area of the branch: The frustum of a cone volume equation, Eq. 3 was used to estimate the volume of the branch: where, r represents the semidiameter of the upper end of the branch, R the semidiameter of the lower end of the branch, h the branch length, and l the generatrix.

Data analysis
To study the spatiotemporal variation in efflux of Dahurian larch and relationships to environmental factors, normality tests and homogeneity of variance tests were carried out on each variable. Subsequently, variance analysis and multiple comparisons were conducted on E and environmental factors at different times and spaces. Correlation analysis was used to ensure the relationship between E and the environmental factors. Regression analysis was used to clarify the quantitative relationships between E and environmental factors, and a quantitative relationship model was established. E and temperature (T) were exponentially fitted as follows, Eq. 4: The temperature sensitivity coefficient (Q 10 ) was calculated according to Eq. 5: All statistical data were analyzed using SPSS 26.0 and R, and figures generated using Origin 2018.

Seasonal variation
Bud burst and leaf expansion did not begin until mid-May, and therefore the experiment of E began in late May. Seasonal dynamics of E in stands of different ages were consistent with single peak types, and E was significantly greater during the growing season than during the dormant season (Fig. 2). As the trees grew rapidly in late May, and vigorously from early June until late July, the efflux values of the three stands gradually increased, and peaked in July (0.7, 1.4 and 1.4 μmol·m -2 ·s -1 , respectively). The E values of stands 31a and 47a in early July were significantly higher than those in the other months (P < 0.05). As temperatures began to decline in mid-August, E gradually decreased, reaching minimum values (0.3, 0.2 and 0.4 μmol·m -2 ·s -1 , respectively) in early October. The average E values increased with increasing age, i.e., stand 27a (0.5 μmol·m -2 ·s -1 ) < stand 31a (0.7 μmol·m -2 ·s -1 ) < stand 47a (0.9 μmol·m -2 ·s -1 ).

Spatial variation Variation in different directions
There were significant E differences in the different directions in the three plantations (P < 0.05) (Fig. 3). The seasonal E variation for all directions peaked once. From May to September, except for June in stand 47a, E on the south side during each month was the largest among the three stands. The mean E of stands 27a, 31a and 47a on the south side were 0.5, 1.0 and 1.1 μmol·m -2 ·s -1 , respectively. Except for August in stand 31a, minimum values occurred in the north, and mean E values were 0.3, 0.7 and 0.5 μmol·m -2 ·s -1 , respectively. In October, the maximum E values of stands 27a, 31a and 47a occurred on the west side (0.3 μmol·m -2 ·s -1 ), the east side (0.3 μmol·m -2 ·s -1 ) and on the south side (0.5 μmol·m -2 ·s -1 ). Minimum values occurred on the east side (0.1 μmol·m -2 ·s -1 ), north side (0.2 μmol·m -2 ·s -1 ) and north side (0.1 μmol·m -2 ·s -1 ). The variations in E in the different directions were such that the south side > east side > west side > north side.

Variation in different canopy heights
There were significant E differences among the three canopy heights in the three stands (P < 0.05) (Fig. 4). Except for July and September in stand 27a, the maximum E in each month occurred in the upper canopy. The mean values of E of the upper canopy in stands 27a, 31a and 47a were 0.5, 1.0 and 1.0 μmol·m -2 ·s -1 , respectively. From May to October, the minimum values occurred in the lower canopy in stands 27a, 31a and 47a (0.3, 0.7 and 0.5 μmol·m -2 ·s -1 , respectively). Overall, E values increased with increasing canopy height.

Effects of environmental factors on the seasonal variation in E
The temperatures of the three stands peaked early July to late July; for stands 27a, 31a and 47a, they were 24.3 °C, 23.0 °C and 21.1 °C, respectively. Beginning in late August, temperatures began to gradually decrease. The minimum temperatures in early October in stands 27a, 31a and 47a were 7.7 °C, 4.8 °C and 4.4 °C, respectively. Temperature differences between months were significant (P < 0.05). Relative humidity of the three plantations increased from May and peaked August to September, for stands 27a, 31a and 47a, was 67.2%, 77.7% and 72.8%, respectively, after which it decreased to low levels in early October (stands 27a, 31a and 47a were 37.0%, 21.9% and 32.0%, respectively). There were significant differences in relative humidity across months (P < 0.05). Seasonal variations in atmospheric pressure and CO 2 levels in the three stands were not obvious. Atmospheric pressures for stands 27a, 31a and 47a were 81.1-81.7 kPa, 80.4-81.4 kPa and 80.8-81.7 kPa, respectively. There was no significant difference in atmospheric pressure between months (P > 0.05). The minimum CO 2 occurred from late August to early September, with values of 433.5 ppm, 420.0 ppm and 417.1 ppm in the three stands, respectively. There were significant differences in CO 2 concentrations across months (P < 0.05) (Fig. S2).
The Dahurian larch began to grow in mid-late May with the vigorous growth from late June to early August. During this time, temperatures and relative humidities were high, providing good hydrothermal conditions for tree respiration and strong metabolic activites. The trees began to lose their leaves in September, stopped growing in late September, and were fully dormant in October. At this time, temperatures and atmospheric relative humidity decreased and E gradually decreased, with the minimum in October. Overall, seasonal E variation was consistent with temperature and relative humidity variations, showing a trend of increasing and then decreasing. However, relationships between E, air pressure and atmospheric CO 2 concentration were not obvious.

Effects of environmental factors on the spatial variation in E (1) Effects of environmental factors on the variation in E in different directions
There were no significant differences in temperature among the three stands across directions (P > 0.05), and the average temperature on the south side in stand 27a was the highest, 16.9 °C. The lowest average temperature occurred on the east side of stand 47a, 14.4 °C. Relative humidity differences between south and north sides were significant (P < 0.05); there were no significant differences in the different directions in stands 27a or 31a. Except for the significant differences in CO 2 levels between the west and north sides in stand 27a (P < 0.05), there was no significant differences in the other stands (P > 0.05). In addition, the atmospheric pressure in the three stands were not significantly different across directions (P > 0.05). Although there was no significant difference in temperature in different directions, temperatures on the south and east sides were slightly higher. Higher temperatures may be due to light conditions. Overall, the spatial variation of E in different directions was consistent with temperature and relative humidity, E on the south and east sides were higher than the other two sides. Temperature and humidity on the south side and east sides were more amenable to growth, providing good environmental conditions for (2) Effects of environmental factors on the variation in E in different canopy heights.
There were no significant differences in temperature among the canopy layers (P > 0.05). The average temperature of the upper canopy in stand 27a was the highest at 17.4 °C, while the lowest occurred in the lower canopy in stand 47a at 14.2 °C. With an increase in canopy height, temperatures gradually increased. There were significant differences in atmospheric relative humidity among the different canopy layers (P < 0.05). Relative humidity in stands 27a and 31a increased with increasing canopy height, while in stand 47a, it was highest in mid canopy. Atmospheric CO 2 in stands 27a and 31a gradually increased with increasing canopy height; in contrast, maximum CO 2 in stand 47a occurred in the lower canopy. CO 2 in the lower canopy was significantly different from that in the upper and middle canopies (P < 0.05). In addition, there was no significant difference in atmospheric pressure among the different canopy heights across the three stands (P > 0.05) (Fig. S4). Except for individual trees, temperatures and relative humidity in the upper and middle canopies were higher than in the lower canopy. The spatial variation of E in different canopy heights was consistent with temperature and atmospheric relative humidity variations; E in the upper and middle canopies were higher than that in the lower canopy. The environmental conditions in the upper and middle canopies were more amenable to growth than those in the lower canopy.

Correlations between E and environmental factors
Correlation analysis showed that E was positively related to temperature (P < 0.01) and relative humidity (P < 0.01). The coefficients between E and temperature in stands 27a, 31a and 47a were 0.51, 0.94 and 0.66, respectively, and those between E and atmospheric relative humidity were 0.53, 0.76 and 0.34, respectively. In addition, there was a significant negative correlation (P < 0.01) between E and CO 2 in stand 27a, with a coefficient of -0.40. However, there was no significant difference between E and CO 2 in stands 31a and 47a (P > 0.05). Efflux was not affected by atmospheric pressure, reflected by low correlation values (0.31 and 0.46) in stands 27a and 31a, respectively. Moreover, there was no correlation between E and atmospheric pressure in stand 47a (P > 0.05). In summary, temperature and relative humidity were the major influencing factors of E (Fig. 5).
There was a significant exponential relationship between E and temperature (P < 0.01), the effect was highest (0.91) in the 31 year-old stand and the lowest (0.80) in the 27 year-old stand (Fig. 6). Similarly, the linear relationship between E and relative humidity in stands 27a, 31a and 47a were distinct (P < 0.05). The effect of relative humidity was highest in the 31 year-old stand (0.57) (Fig. 7).

Seasonal and spatial variation in the temperature sensitivity coefficient (Q 10 )
The temporal trends of temperature and Q 10 were dissimilar (Fig. 8). The monthly mean temperatures increased beginning in May (16.7 °C) and peaked in July (22.2 °C), after which they decreased to their lowest value (4.8 °C) in October. In contrast, Q 10 decreased beginning in May (2.9), reaching its minimum in July (1.1), after which it increased to peak in October (4.0).

Fig. 5
Correlation coefficients between CO 2 efflux and environmental factors for branches and leaves of 27 (left)-, 31 (middle)-and 47 year-old (right) old Dahurian larch stands. Correlation coefficients correspond to diagonal values; the larger the circular area, the greater the correlation coefficient; blue is a positive correlation, red a negative one; P values of correlation coefficients greater than 0.05 are marked as " × " Exponential fitting of E and temperature at different positions was conducted to evaluate the mechanism through which temperature affects E and to identify a quick and convenient way to estimate E for each position (Table 2). A relationship was found between E and temperature at each position for the three different stands (R 2 > 0.72). The most sensitive positions of E to temperature in stand 27a (Q 10 = 3.11) and stand 31a (Q 10 = 3.44) were the east side of the mid canopy. Conversely, the most sensitive position in stand 47a was the west side of the mid canopy (Q 10 = 4.24). In summary, 75.8-92.2%, 60.3-94.6% and 56.5-79.8% of the seasonal variation in efflux at different positions could be explained by temperature in the 27-, 31-and 47 year-old stands, respectively.

Temporal variation in efflux
Efflux values of the different stands ranged from 0.1 to 4.7 μmol·m -2 ·s -1 . Our results are larger than those of branch efflux measured for Larix kaempferi (Xia 2017) and those for L. gmelinii (Jiang 2003). It is speculated that this might be due to the different habitats of the species and the large differences in environmental temperatures. In addition, different instruments were used in the experiments. LI-8100A works to maintain a pressure equilibrium between the inside of the chamber and the external environment during measurements to avoid a decrease in respiration over time due to high chamber pressure. Zhou et al. (2010) showed that the pattern of diurnal respiration of coniferous forests exhibited a single peak.  found that the diurnal variation in leaf respiration first increased and subsequently decreased. In this study, diurnal E variations was consistent with variations in temperature and showed a single peak, which is consistent with previous results. These indicate that the diurnal variation of CO 2 efflux of branches and leaves is affected by temperature. The response of CO 2 efflux to temperature is that E increases with increasing temperature and decreases with decreasing temperature. The response time of E to air temperature change was under 2 h. Lavigne (1987) found that sap flow could mediate the relationship between CO 2 efflux and temperature, and the response time of E to temperature changes may be related to sap flow (Zhao et al. 2018).
In terms of seasonal variation, studies found that leaf efflux exhibits clear seasonal dynamics, and a peak generally Fig. 6 Relationships between E and temperature (T) in 27-, 31-, 47 year-old stands occurs in June (Xu and Griffin 2008;Catoni et al. 2013;. Other studies have found that branch efflux also exhibits seasonal dynamics with a peak often occurring early in the growing season, and higher than that in the dormant season (Shi et al. 2010;Han et al. 2015;Xia 2017). The results of this study also indicated obvious seasonal variation. The maximum temperature occurred in July, and the minimum in October. Seasonal variation in E was mainly affected by temperature. The peak E in the three stands occurred in early July, while E was lowest in October. Energy and carbon skeletons generated by respiration can provide guarantee for physiological activities of plants. During the growing season, plant tissues are highly physiologically active and need sufficient substrates, necessitating greater respiration to supply the carbon skeletons and energy for growth (Sun et al. 2013;Han et al. 2015). The temporal variation in E of L. gmelinii var. principis-rupprechtii plantations can be further clarified by improving the research methods and measuring E separately.

Spatial variation in E
In this study, direction had significant effects on E, and varied such that south > east > west > north. Due to the sun, leaves facing different directions received different quantities of light, resulting in differences in temperature and humidity, and the structure and physiological characteristics of the leaves also changed (He et al. 2010). Leaves on the south side received sufficient light, resulting in a high photosynthetic rate and more photo assimilates, thereby providing sufficient substrates for respiration on the southern sides of the trees. In contrast, leaves on the north side did not receive sufficient light, resulting in low photosynthetic rates and low amounts of photo assimilate. Thus, respiration may be limited by the supply of substrates. Acosta et al. (2011) and Asao et al. (2015) found that the CO 2 efflux of branches differed in different directions, and on the eastern side was generally higher than on the western. This study showed that there are differences in environmental factors across the different directions but are not significant. Although there is a strong correlation between E and temperature, other factors may be involved in the E response to direction, which could mask the effect of temperature.
With regards to the vertical position, studies have found that the CO 2 efflux in the upper canopy is greater than in the lower canopy, i.e., the higher in the canopy, the greater the respiration Acosta et al. 2011;Weerasinghe et al. 2014). In this study, E values increase with canopy height. The higher respiratory rate in the upper canopy may be due to the higher mitochondrial density per cell needed to support higher metabolism (Tissue et al. 2002). In addition to the high mitochondria density, respiratory rate is also affected by the supply of respiratory substrates and energy demand (Lambers et al. 2008). Compared with the middle and lower portions of the canopies, the upper canopy receives direct sunlight, accumulates more sugar and starch through photosynthesis, and provides adequate substrates for respiration (Azcón-Bieto and Osmond 1983). A higher respiratory rate in the upper canopy reflects higher energy needs that may stem from protein repair, maintenance of solute gradients, and carbohydrate loading (Ryan 1991;Amthor 2000;Bouma 2005).

Environmental factors influencing variations in E
This study found that efflux is affected by temperature and relative humidity. Temperature directly affects plant respiration but indirectly affects respiration by driving changes in nitrogen content, leaf area, and soluble sugar concentrations in tissues (Lee et al. 2005). Respiration involves a complex series of enzymatic reactions that are regulated by temperature. When temperatures are at low level (e.g. 5 °C), respiration is limited by enzymatic activities. As temperatures rise (e.g. 25 °C), the restriction of enzyme temperatures is alleviated, and substrate supply becomes the main factor limiting respiration (Atkin and Tjoelker 2003). Plants need to accumulate respiratory substrates through photosynthesis, so respiration is also affected. Further, relative humidity affects the rate of electron transport via the water content of plant cells. Under low humid conditions, the electron transfer efficiency of enzymatic reactions is low, resulting in a lower respiratory rate; in contrast, adequate moisture promotes electron transport and accelerates cellular respiration (Zhao et al. 2018;Zhao 2020). However, light can inhibit plant respiration (Parnik et al. 2010;Ayub et al. 2014). For example, Sun et al. (2015) found that daily respiration decreased with increasing photosynthetic photon flux density, and a small amount of light can also inhibit respiration. However, Wang and Mao (2007) suggested that light was not the main factor affecting respiration. In summary, respiration is affected by various environmental factors. Therefore, additional comprehensive studies are needed to explore the response of efflux to environmental factors.

Seasonal and spatial variation in temperature sensitivity coefficient (Q 10 )
Q 10 is an important indicator for studying plant respiration and can be used to evaluate the carbon budget dynamics of forest ecosystems. Among tree species, leaf and wood tissues have different sensitivities to temperature. Brito et al. (2013) found that the temperature sensitivity coefficient of leaves decreased with increasing temperature, while the temperature sensitivity of woody tissues (such as branches) increased with increasing temperature. This study showed that the seasonal variation in Q 10 was inversely related to changes in temperature, and Q 10 values in the early and late growing seasons were higher than those in the vigorous growth period. Our results are consistent with those of Araki et al. (2017) but differ from those of Wan et al. (2019). The temperature dependence of Q 10 is related to enzyme activities at low temperatures and substrate limitations at high temperatures. When respiration is limited by enzyme activity, Q 10 is higher than when respiration is limited by substrate supply. With increasing temperature, the restriction conditions change from enzyme activity to substrate supply, and the Q 10 value decreases (Atkin and Tjoelker 2003).
Nitrogen is a component of respiration-related enzymes, and soluble sugars are important respiratory substrates. Their seasonal variations may affect the response of respiration to temperature, resulting in seasonal variations in Q 10 (Wan et al. 2019). The variation in Q 10 with height generally occurs when the temperature sensitivity in higher in the canopies is greater than in lower canopies. Moreover, Q 10 values are positively correlated with substrate concentration (Atkin and Tjoelker 2003). The upper and middle canopies receive ample light and accumulate more photosynthetic products which provide sufficient substrates for respiration, resulting in increased Q 10 values.

Conclusions
The efflux of Dahurian larch shows obvious temporal and spatial variations. Diurnal and seasonal variations exhibited a single peak in the three different aged forests. On a temporal scale, efflux increased with stand age, such that stand 27a < 31a < 47a. For spatial variation, efflux values of the three stands showed a regularity across directions. Efflux values on the south and east sides were higher. Overall, efflux values were such that south > east > west > north. Variations among the different canopy layers showed that efflux increased with height. In addition, temperature and relative humidity were the main environmental factors affecting efflux, and increased with increasing temperature and relative humidity. The seasonal variation in Q 10 was opposite that in temperature. Q 10 increased with decreasing temperature. On the spatial scale, the maximum Q 10 occurred in the mid canopy in the three stands. With the efflux and temperature models in different locations, efflux variation can be better estimated at each site. Further research is needed to explore the efflux response to environmental factors.
Author contributions LL and XW have contributed equally to this work and share co-first authorship. LL and XW designed the study, performed the experiments and wrote the paper. ZJ reviewed and edited the manuscript.
Open Access This 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/ licen ses/ by/4. 0/.