Diurnal and Seasonal Change in Stem Respiration of Larix principis-rupprechtii Trees, Northern China

Stem respiration is a critical and uncertain component of ecosystem carbon cycle. Few studies reported diurnal change in stem respiration as well as its linkage with climate. In this study, we investigated the diurnal and seasonal change in stem respiration and its linkage with environmental factors, in larch plantations of northern China from 2010 to 2012. The stem respiration per unit surface area (RS) showed clear diurnal cycles, ranging from 1.65±0.10 to 2.69±0.15 µmol m−2 s−1, increased after 6∶00, peaked at 15∶00 and then decreased. Both stem temperature and air temperature show similar diurnal pattern, while the diurnal pattern of air relative humidity is just the opposite to Rs. Similar to the diurnal cycles, seasonal change in RS followed the pattern of stem temperature. RS increased from May (1.28±0.07 µmol m−2 s−1) when the stem temperature was relatively low and peaked in July (3.02±0.10 µmol m−2 s−1) when the stem temperature was also the highest. Further regression analyses show that RS exponentially increases with increasing temperature, and the Q10 of Rs at mid daytime (1.97±0.17 at 12∶00 and 1.96±0.10 at 15∶00) is significantly lower than that of mid nighttime (2.60±0.14 at 00∶00 and 2.71±0.25 at 03∶00) Q10. This result not only implies that Rs is more sensitive to night than day warming, but also highlights that temperature responses of Rs estimated by only daytime measurement can lead to underestimated stem respiration increase under global warming, especially considering that temperature increase is faster during nighttime.


Introduction
Rising atmospheric carbon dioxide (CO 2 ) is considered to have significant impacts on the climate system [1], which has triggered strong scientific interests in understanding the global carbon cycle. Forests play a key role in the global carbon cycle. They cover approximately one-third of the earth's land surface, and store about 861666 Pg of the total carbon [2]. Current terrestrial carbon sink has also been suggested to be mainly contributed by the forest sink [2]. Accordingly, accurate information on processes related to forest carbon cycle is essential to predict future evolution of the global carbon cycle and climate change. As a major pathway of carbon loss from terrestrial ecosystems, ecosystem respiration is critical to regulating forest ecosystem carbon fluxes and thus important to forest carbon balance. Ecosystem respiration is composed of two dominant fluxes, (i) soil respiration including heterotrophic respiration of decomposing microbes, respiration of plant roots and soil fauna, and (ii) above-ground respiration of plant woody tissues and leaves. Compared with soil respiration [3][4][5], our understanding of the linkage between above-ground respiration and climate is very limited [6,7].
Temperature sensitivity of stem respiration is usually expressed in terms of Q 10 (the rate of change in respiration resulting from a 10uC increase in temperature). Numerous studies on temperature sensitivity of stem respiration have been conducted across different forest types of the world and reported different Q 10 values of stem respiration for different forests, varying from 1.00 to 6.40 [26,[29][30][31][32]. It should be noted, however, that most of these previous studies estimated Q 10 values based on the measurement of daytime stem respiration [33], and few studies measured diurnal change in stem respiration as well as its linkage with climate [34]. Since stem respiration is also influenced by other environmental and physiological processes [16,23,[35][36][37][38][39], such as photosynthesis that occurs only during the daytime, it is possible that stem respiration responds to temperature changes in daytime and nighttime differently. Furthermore, both observations and model projection have showed that global warming is faster during the nighttime than that during the daytime [1]. Thus, understanding the possible differential responses of stem respiration to day and night warming will be helpful to improve the projection of future carbon cycle evolution as well as its feedback to climate.
In this study, we have conducted field measurement to investigate the diurnal and seasonal change in stem respiration and its linkage with environmental factors, in larch plantations of northern China since 2010. The primary object of this paper is to test the hypothesis that temperature sensitivity of stem respiration is different during daytime and nighttime.

Study Site and Experimental Design
This study was conducted at Saihanba ecological station (42u24.7239N, 117u14.8449E, 1505 m a.s.l) of Peking University, situated in Saihanba National Forest Park, Hebei Province (Fig. 1). Saihanba has a mean annual precipitation of approximately 450 mm, 70% of which occurs from June to August, and mean annual temperature of 21.4uC [40] with a long cold winter and a short growing season (May-September). The soils are predominantly sand. Soil bulk density is 1.47 g cm 23 , C:N ratio is 8.960.3, and soil pH (soil:water, 1:2.5) is 6.360.2 [41].
The experiment was carried out in three 20620 m plots located within a 45-year-old larch plantation (Larix principis-rupprechtii). The topography of the plots is nearly flat and the stem density is 870648 stem?ha 21 with an average diameter at breast height (DBH) of 19.962.8 cm and an average height of 15.861.6 m. Two larch trees were chosen randomly in each plot, and all together 6 trees were selected in the study area with an average DBH of 2062 cm and an average height of 1661.5 m. Although the experiment plots and individuals were very homogeneous, it should be reminded that the limited sample size (6 trees in total) might introduce biases originated from inter-individual differences and the effect of micro-topography. In this study, 3-hourly stem respiration measurements were made for a whole day, twice a month in growing season from 2010 to 2012, using a LI-6400-09 (Li-Cor, Lincoln, Nebraska, USA). In order to capture the CO 2 released by stems, a technique called horizontally oriented soil chamber (HOSC) [12,42] was exploited: the CO 2 chamber (9.9 cm in diameter) was connected to stem collars (10.1 cm in diameter), which completely enclosed a 10.1 cm segment of the tree stem at 1.3 m above ground and were fixed tightly onto the stems with nylon straps. To ensure an airtight seal between stem collars and stem surfaces, loose barks at two ends of the enclosed stem segment, which might leak air, were removed at first. Then the collars made of polyvinyl chloride (PVC) pipe were polished to fit the curvature of the stem surface. Finally the small gaps between the collars and stem surfaces were sealed completely with silicon sealant. Meanwhile, stem temperatures were measured with copper-constantan thermocouples at the depth of 5 mm from the stem surface with tree bark.

Data Analyses
To eliminate the influence of plant size, the measured stem respiration was firstly normalized by the surface area enclosed, which was calculated using the following equation [43]  A~p where A is the stem surface area enclosed by the collar (m 2 ), D 1 is the diameter of the chamber (m), D 2 is the diameter of the enclosed stem segment (m). Then the stem respiration per unit surface area R S (mmol?m 22 ?s 21 ) should be the measurement results divided by the area A for each plant.
The relationship between stem respiration and corresponding stem temperature can be described by an exponential function where T is the measured stem temperature (uC); R S is the stem respiration per unit surface area (mmol?m 22 ?s 21 ) at temperature T; R 0 is the potential stem respiration rate at 0uC and b is a fitting parameter, which indicates the temperature sensitivity of respiration [44]. The temperature sensitivity is often expressed by Q 10 , which describes the proportional change in stem respiration rate for a 10uC increase in sapwood temperature). According to Eq.2, the Q 10 values can be calculated as For each plant individual, three-hourly R 0 and b were acquired by fitting stem respiration (R) and stem temperature (T) measured throughout the sampling period to Eq.2 and three-hourly Q 10 values were calculated according to Eq.3. Then they are averaged for all six larch individuals. All the statistical analyses were performed in PASW statistic 18 (SPSS Inc., Chicago, IL, USA).

Diurnal Variation
The measured stem respiration per unit surface area (R S ) and environmental factors showed clear diurnal cycles (Fig. 2), averaged over the whole sampling period. R S , ranging from 1.6560.10 to 2.6960.15 mmol m 22 s 21 , increased after 6:00, peaked at 15:00 and then decreased. Both stem temperature and air temperature showed similar diurnal pattern. Nevertheless, stem temperature experienced a plateau after mid-day (12:00-15:00) and then decreased more quickly than R S . As shown in Fig. 2A and C, the stem temperature values were comparable between 09:00 and 18:00 but the R S value was much larger at 18:00. Air temperature had a similar fast afternoon decrease pattern as stem temperature did, but the amplitude of air temperature diurnal change (12.75uC) was larger than that of stem temperature (9.02uC). The diurnal pattern of air relative humidity was just the opposite to that of air temperature (Fig. 2B and D), ranging from 46.6462.21% at 12:00 to 89.4562.13% at 3:00. Figure 3 illustrated the seasonal changes of R S , air relative humidity, stem temperature, and air temperature from May to September. Similar to the diurnal cycles, R S followed the pattern of stem temperature ( Fig. 3A and C). R S increased from May (1.2860.07 mmol m 22 s 21 ) when the stem temperature was relatively low (13.7060.47uC) and peaked in July (3.0260.10 mmol m 22 s 21 ) when the stem temperature was also the highest (17.7360.30uC). Both of them decreased afterwards to the lowest point in September, with R S as 1.1960.05 mmol m 22 s 21 and stem temperature as 9.1360.56uC. It is also noteworthy that the R S values did not differ very much between the start and the end of the growing season, while there was a significant gap between the stem temperature values (near 14uC in May but around 9uC in September). Similarly, the air temperature reached its peak value in June and July (Fig. 3D), and the minimum value occurred in September. Consistent with diurnal changes, seasonal maximum air temperatures were higher than maximum stem temperatures in June and July, meanwhile, seasonal minimum stem temperature was 8.80uC, lower than that of air temperature (9.13uC) in September. The seasonal pattern of air relative humidity was no longer the opposite to that of air temperature (Fig. 3B), which increased in early growing season, decreased a little in July, peaked in August and then dropped in September.

Diurnal Change in Q 10
To gain further understanding of how environmental factors influence stem respiration activity, the stem respiration rates and temperature measured in the sampling period were fitted to Eq.2 ( Fig. 4A and B) and stem respiration rates were also linearly regressed against the air relatively humidity (Fig. 4C). In general, R S showed a good exponential relationship with both stem temperature (Fig. 4A, R 2 = 0.47, P,0.001) and air temperature (Fig. 4B, R 2 = 0.39, P,0.001). There was no good linear relationship between R S and air relative humidity (Fig. 4C, R 2 = 0.00).
In order to investigate diurnal variations of temperate sensitivity of R S , Q 10 values were further calculated based on the seasonal variation in R S for each time during one day. There was statistically significant difference between daytime (1.9760.17 at 12:00 and 1.9660.10 at 15:00) and nighttime (2.6060.14 at 00:00 and 2.7160.25 at 03:00) Q 10 (Fig. 5). Q 10 values in other time intervals fell in between and were not significantly different from each other.

Discussion
The magnitude of our R S values (0.33-6.59 mmol m 22 s 21 ) is similar with previous studies on mature conifer forests. For example, Wang et al. [43] found that stem respiration rates in a 33-year-old larch forest varied from approximately 0.9 mmol m 22 s 21 to 6.6 mmol m 22 s 21 in June, 2001 and Acosta et al. [8] documented that the R S range of a 22-year-old Norway spruce forest stand during the growing season from 1999 to 2002 was 0.34-6.52 mmol m 22 s 21 . Meanwhile, the mean stem respiration (2.15 mmol m 22 s 21 ) was lower than that of soil respiration rate (3.22 mmol m 22 s 21 ) [41] at the same plot. Bolstad et al. [45] showed that the stem respiration was lower than the soil respiration which was typically more than 60% of total ecosystem respiration during the growing season while Clinton et al. [46] showed that the mean stem CO 2 efflux (2.6060.17 mmol m 21 s 22 ) was slightly higher than that of soil CO 2 efflux (2.5360.11 mmol m 21 s 22 ).
Stem respiration rates can respond to temperature changes and plant activities like photosynthesis, plant growth, etc. [47,48]. Our results show that variations of stem respiration rates in larch forests were largely influenced by diurnal and seasonal changes of stem temperature. During the study period, the maximum of the stem respiration occurred in the afternoon while the minimum occurred in the early morning within one day and stem respiration rates peaked in July in growing season, which is consistent with previous studies [9,18,23,31,42,49]. For example, Zha et al. [9] found that stem respiration of Scots pine peaked at around 16 h and was highest in July. Acosta et al. [8] indicated that stem respiration of Norway spruce reached maximum between 13 h and 16 h and the highest rate occurred in June and July. Zhu et al. [23] suggested that stem respiration of Schima superba also followed a similar diurnal pattern, reaching the highest in the afternoon and the lowest at about 8:00 in the early morning.
Nevertheless, stem temperature can't fully explain all of the variations of R S [9]. In our study, the R S values of afternoon (18:00) and late growing season (Aug) were higher than those in  the morning (09:00) and early growing season (May) while the stem temperature were comparable ( Fig. 2A and C, Fig. 3A and C). This phenomenon suggests that plant activities like photosynthesis and cambium activity probably play an important role in regulating stem respiration changes. Martin et al. [50] found that when temperature and transpiration are constant, R S appears to be positively correlated with substrate supply. The diurnal change of respiratory substrate, supplied by photosynthesis, may also influence the respiration rates [48,51]. At seasonal scale, plant growth activities can't be ignored. The stem respiration mainly consists of maintenance respiration and growth respiration [52]. Maintenance respiration varies primarily with changes in temperature and is also reported to increase with relative growth rate [37,53]. Meanwhile, growth respiration is controlled by the timing and magnitude of plant growth [54]. That is to say, the stem respiration varies throughout the growing season, following not only the change of temperature, but also the change of phenology and environmental factors that control growth.
Accurate understanding of temperature response of respiration is critical in estimating global carbon balance and its response to current climate change. Our results show that Q 10 values of Rs vary from 1.96 to 2.71, which are within the range reported by previous studies [9,17,31,34,42,55,56]. In growing season average diurnal cycle, Q 10 values were lowest in mid-day (12:00-15:00), which may be partly explained by the acclimation of respiration to rising temperature. Both theory and observations have suggested a decline temperature sensitivity of rates of respiratory CO 2 efflux from plants [52,57,58] and soils [5,59]. For example, Tjoelker et al. [57] reported that Q 10 value of foliar respiration decline by 0.04 in response to 1uC increase in mean ambient temperature. In deed, highest stem temperature is observed during the mid-day. Another possible explanation of the suppressed temperature sensitivity of stem respiration in mid-day (12:00-15:00) may be midday depression of photosynthesis, particularly during summer with stem temperature approaching 30uC and air temperature above 30uC (Fig. 4A&B). High midday temperature is considered to be able to induce stomata closure and photosynthesis depression in water-limited regions by both observations (pine forest in Canary Islands) [60] and theoretical models. Reduced stomatal conductance and photosynthesis rates in midday during summer may further decline Rs, and thus influencing Q 10 values derived from seasonal variation of R S . Often in models [14,31,61], Q 10 is set to be a constant value of 2, similar to the midday values and lower than the nighttime values from our study. This can lead to underestimated stem respiration increase under global warming, especially considering that temperature increase is faster during nighttime [1].
Vegetation activities have been shown to respond negatively to nighttime temperature increase in cold and mesic regions [65], probably due to increased carbon loss through respiration. Combined with stronger nighttime warming [1,62,63], our results imply that the carbon loss through respiration might increase more than former model projections [14,31,61], and might further cancel out the increased photosynthesis driven by daytime warming in those areas. In contrast, ecosystems in arid and semi-arid regions are thought to respond to night time temperature change in a more complex way [65]. A manipulative experiment study in a temperate steppe ecosystem in north China reported that daytime warming induced reduction in gross ecosystem productivity (GEP), and night-time warming stimulated photosynthesis and GEP in the following day because enhanced respiration drew down the leaf carbohydrates concentration [64]. With higher stem respiration sensitivity at night, the stimulation effect might be strengthened in the future, while it is also possible that the carbon loss through enhanced nighttime respiration goes up even faster and cancels out the stimulating effect. Thus, more experimental researches as well as modelling efforts are necessary to accurately quantify the temperature sensitivity of stem respiration and to better address its implications on future vegetation dynamics.

Conclusion
Temperature responses of rates of respiratory CO 2 efflux from plants and soils are generally modelled using exponential functions with a constant Q 10 near 2.0, similar to the midday values and lower than the nighttime values from our study. This result has important implications for the predictions of forest responses to warming. Current carbon cycle models must consider diurnal change in temperature sensitivity of Rs to accurately predict ecosystem C cycling under climate warming. In the future, additional experiments with larger sample size need to be performed in other ecosystems in order to draw a more generalized conclusion and to further address detailed mechanisms responsible for diurnal change in Q 10 of Rs.