The effect of thermal acclimation on increase in leaf or canopy R with rising temperature is relatively high in boreal and temperate trees Reich et al. 2016 , moderate in a temperate evergreen conifer i.e., approximately 60% in this study , and lower in tropical trees and lianas by Slot et al. 2014 , although the value presented by Reich et al. 2016 was not based on annual estimates and the estimation methods differ among studies. This suggests that thermal acclimation capacity may vary with plant functional types and biomes. Therefore, clarifying whether thermal acclimation capacity of plant R is various or convergent across time and space Slot and Kitajima 2015, Vanderwel et al. 2015 is important for implementing the acclimation into global terrestrial carbon cycle models.
The simulation in this study was conducted based on the assumption that the seasonal temperature acclimation capacity of leaf R observed for hinoki cypress, i.e., the negative correlations of R20 and Q10 with ambient air temperature Figure 2.9 , would continue unchanged in future warmer climatic conditions. Recently, Aspinwall et al. 2016 found that acclimation of leaf R to seasonal temperature changes was equivalent to acclimation to experimental warming 3 °C above ambient of a similar magnitude for Eucalyptus tereticornis trees growing in the field. Similar results were demonstrated for boreal and temperate trees, based on a relatively long‐term warming experiment in the field Reich et al. 2016, Wei et al. 2017 . These novel findings support that the assumption used in this study might be reasonable. Moreover, tree leaves developed at a new warmer temperature generally have higher capacity to acclimate to the new temperature relative to pre‐existing leaves Atkin and Tjoelker 2003, Loveys et al. 2003, Campbell et al. 2007 . Longevity of hinoki cypress leaves are approximately 4‒6 years Miyamoto et al. 2013 and climatic warming will bring a gradual increase in temperature year by year. Therefore, it is likely that the acclimation capacity may gradually change and, consequently, canopy R may be downregulated to a further degree compared with the predictions in the current simulation.
4.4.2 Effect of thermal acclimation on prediction of stem maintenance respiration
The present study found that stem Q10 varied seasonally and negatively correlated with ambient air temperature Figure 4.4 . This result suggests that stem maintenance respiration can acclimate to a seasonal temperature change. Consequently, the simulation model of stem CO2 efflux incorporating seasonally variable Q10 was able to be developed i.e., Equation 4.14 . In comparison with the case of
canopy R, effect of the thermal acclimation was less distinct on seasonal patterns of stem Rm Figure 4.7 and estimates of annual total stem Rm under the present climatic condition Table 4.4 . Thus, acclimation capacity to a seasonal temperature change might be lower in stem Rm than in leaf R. This might be partly attributed to the slope values in the relationship between leaf Q10 or stem Q10 and ambient air temperature i.e., 0.034 in leaf vs. 0.027 in stem; see Equations 4.8 and 4.13 . Nevertheless, increases in monthly Rm in response to climate warming were mitigated during summer by the acclimation. Consequently, increases in annual total stem Rm in the 50‐year‐old hinoki cypress stand under the future climate conditions were reduced with acclimation relative to without acclimation
18% and 26% based on the RCP2.6 and RCP8.5, respectively; Table 4.4 .
Evidence showing thermal acclimation in woody tissue respiration might be very rare. Recently, Drake et al. 2016 reported that branch wood respiration at 15 °C in E. tereticornis trees was reduced by ~23% with experimental warming 3 °C above ambient , whereas the rate of leaf R at 15 °C was reduced by ~30% with warming. These results suggest that woody tissue R can acclimate to warming and its acclimation capacity is lower compared to leaf R, which agree with findings in this study.
Additionally, the negative correlation of stem Q10 with ambient air temperature found in this study is consistent with leaf Q10 Tjoelker et al. 2001, Atkin and Tjoelker 2003, this study and Q10 of R in the entire aboveground parts of hinoki cypress trees Paembonan et al. 1991, Yokota and Hagihara 1996b . In contrast, several studies on stem CO2 efflux have observed no clear seasonality in stem Q10 Ceschia et al. 2002, Damesin et al. 2002, Wieser et al. 2009, Tarvainen et al. 2014, Tu et al. 2017 . This contradiction may be resulted from stem respiration consists of the maintenance and growth components Ryan 1990, Sprugel 1990 , and from the fact that stem CO2 efflux does not reflect stem respiration directly due to complex CO2 pathways within tree stems such as storage, transport, and recycling Pfanz et al. 2002, Teskey et al. 2008, Trumbore et al. 2013, see Chapter 3 for detail .
On the other hand, it was found that stem Q10, which was obtained by re‐analyzing short‐term temperature responses, did not vary within or between trees data not shown . This is consistent with the results in relation to the long‐term temperature sensitivity of Es long‐term Q10, Figure 3.7 and with Tarvainen et al. 2014 . Incidentally, a long‐term Q10 of Es calculated from using all data set was 1.72 i.e., h in Equation 4.11, Figure 4.3a, Table 4.3 . The growth respiration coefficient see Equation 1.3 was 0.38 in hinoki cypress stems, which was calculated from data presented in Tables 3.5 and 3.6.
Such information might be useful for carbon cycle models that use the traditional two‐functional model.
However, it is still important to elucidate detailed respiratory behavior of woody tissue respiration, such as seasonality in stem Q10 and thermal acclimation capacity.
4.4.3 Effect of vertical variation in leaf respiration on canopy respiration
In the present study, the single big leaf model resulted in overestimation of the canopy R 133 % compared with the method considering detailed vertical variations in leaf R and leaf area. The considerable difference in the two estimates suggest that it is difficult to determine which leaf R is appropriate as the representative value within the canopy in the big leaf model. For example, even the average of R20 derived from shoots at various heights and across seasons n 88 , which was used as the representative in the current calculation, led to the overestimation. The most likely position suitable for the representative leaf R might be the middle of the canopy where leaves are mostly concentrated. However, in this study, R20 at the middle point of the canopy z 2.8 m led to underestimation of canopy R 86 % . Instead, R20 at slightly higher position z 2.4 m was found to be most appropriate to calculate canopy R using the single big leaf model.
In the hinoki cypress canopy, sunlit and shaded leaves contributed almost evenly to the total leaf area, whereas they contributed 69 % and 31 % to the annual canopy R, respectively. Thus, although the two‐big‐leaf model may estimate canopy R better than the single big leaf model, it potentially has a similar problem. Specifically, the most appropriate height of leaf R to calculate respiration in upper‐sunlit and lower‐shaded canopies were z 1.9 m and z 3.6 m, respectively. However, in general, it may be difficult to detect such heights. The Hozumi‒Yoda model Kira et al. 1969, Yoda 1971 , which uses leaf area‒light‒leaf respiration relationships, could be also applicable to forests with more complex canopy structure such as tropical rain forests Yoda 1983 . Therefore, although calculation procedures become more complicated, the spatial upscaling method that accounts for vertical variations in leaf R and leaf area might be useful for estimating more accurately whole canopy respiration for various forest types.
Annual total of nighttime canopy R in the 10‐year‐old hinoki cypress stand was estimated to be 3.72 Mg C ha 1 year 1 in 2012 Table 4.1 . When annual canopy R without distinguishing day and night
was re‐calculated assuming that daytime and nighttime leaf R are identical at the same temperature i.e., without light inhibition of leaf R , then 9.90 Mg C ha 1 year 1 was obtained. Comparing this value with other literature values in hinoki cypress canopy, the value in this study is slightly larger than 8.79 Mg C ha 1 year 1 estimated for a 17‐year‐old stand based on seasonal and vertical measurements of leaf R Hagihara and Hozumi 1977 . Additionally, Tadaki et al. 1966 showed 8.53 Mg C ha 1 year 1 for a 45‐year‐old unthinned stand on the basis of leaf mass in the stand and respiration of seedlings reported by Negisi and Satoo 1961 , and Saito 1974 provided a much lower value, 2.22 Mg C ha 1 year 1, for a 40‐year‐old stand by using leaf R data obtained in other stand by Oohata et al. 1971 . Note that because these literature values were expressed as unit of dry matter, they were converted to carbon base assuming a conversion factor of 0.614 Mg of dry matter per Mg of CO2 Tadaki et al. 1966, Hagihara and Hozumi 1991 . Surprisingly, although the calculation method differs, these three values Tadaki et al. 1966 , Hagihara and Hozumi 1977 and this study, except for Saito 1974 appear to be almost comparable and may support the age‐related change in canopy R proposed by Kira and Shidei 1967 . However, studies on canopy R in hinoki cypress stands at various ages are required to discuss the age‐related change in canopy R.