Responses of Intrinsic Water-use Efficiency and Tree Growth to Climate Change in Semi-Arid Areas of North China

Tree-level intrinsic water-use efficiency (iWUE) is derived from the tree-ring 13C isotope composition (δ13C) and is an important indicator of the adaptability for trees to climate change. However, there is still uncertainty regarding the relationship between long-term forest ecosystem carbon sequestration capacity and iWUE. To determine whether elevated atmospheric CO2 concentration (Ca) increase iWUE and tree growth (basal area increment, BAI), dendrochronological methods and stable isotope analyses were used to examine annual changes in the tree-ring width and carbon isotope composition (δ13C) of Platycladus orientalis in northern China. The iWUE derived from δ13C has increased significantly (p < 0.01). Long-term iWUE trend was largely and positively driven by the elevated atmospheric CO2 concentration and temperature. We observed a general increase in averaged BAI, which had significant positive correlation with iWUE (R2 = 0.3186, p < 0.01). Increases in iWUE indeed translated into enhanced P. orientalis growth in semi-arid areas of northern China. Elevated atmospheric CO2 concentration significantly (p < 0.01) stimulated P. orientalis biomass accumulation when Ca was less than approximately 320 ppm in the early phase; however, this effect was not pronounced when Ca exceeded 320 ppm.

over short time periods in growth chambers with seedlings or young plants 7,13,14 . The long-term responses of forests to gradually changing environment are complicated by factors that affect several fluxes at the same time, and by complex interactions among the factors 15,16 . For example, the direct response of photosynthesis to an increase in atmospheric CO 2 concentration will require additional N and other nutrients, and the production response will therefore decline as nutrient limitation progresses. Results derived from short-term experiments may not represent longer-term responses of plants to multiple factors in the natural environment 4 . Compared with short-term experiments, tree ring information provides a more reliable summary of physiological responses to environmental variations over time and can be used to predict the behavior of mature trees [17][18][19][20][21] . Studies of tree-ring stable carbon isotopes at a global scale have found variable tree responses to climate change, reflecting the complex relationship between environmental variability and tree growth, and suggest that adaptation strategies vary widely across species and sites [22][23][24][25] . Our knowledge of the long-term relationship between iWUE and tree growth remains incomplete. Tree rings are ideal proxies for assessing changes in physiology and growth over time and under different environmental conditions.
The Yanshan Mountains of northern China, between the Mongolia Plateau and the North China Plain, are forested and dominated by P. orientalis. This area plays a significant role in the global carbon cycle due to its typical monsoon climate. Platycladus orientalis, an evergreen coniferous tree species that is endemic to China, is the main tree species used for afforestation in the semi-arid areas of northern China. Climate change research in China over the last century has focused on the importance of the adaptation strategies and productivity of typical evergreen conifers in the semi-arid region of northern China 13 . In this study, we examined annual changes in carbon isotope composition and tree-ring width, investigating trends in tree-ring carbon discrimination (Δ 13 C) and iWUE over the last century in P. orientalis of the Yanshan Mountains. The specific goals of the study were to (1) determine whether tree-level iWUE increased over time (i.e., across subsequent calendar years) and whether that translated into tree growth and (2) assess how iWUE and tree growth (as measured by the basal area increment [BAI]) responded to changes in atmospheric CO 2 concentrations, annual temperature, and precipitation.

Materials and Methods
Study sites, tree sampling, and analysis. The study was carried out in the Hongmenchuan watershed (40°26′N, 117°10′E) of northern China, which is an important ecological refuge for Beijing and more than two million people. The Hongmenchuan watershed, located in the transition zone of the North China Plain and the Mongolia Plateau, has a complex geology and belongs to the Yanshan Mountain range (200-1200 m a.s.l.). The region is characterized by a sub-humid, warm and temperate continental monsoon climate. The vegetation in the Hongmenchuan watershed is natural forests with no forest management, dominated by P. orientalis with a density of 739 trees per ha, along with the co-dominant tree species Pinus tabulaeformis, with a density of 116 trees per ha; these data were obtained through sample plot investigation (data not shown).
The detailed information for each plot is shown in Table 1 26 . P. orientalis is mainly distributed in areas with an altitude of 200 m-600 m, on shady and sunny slopes 27 . Therefore, we selected 10 typical plots with altitudes ranging from 332 m to 562 m located on each slope aspect. Within each plot, four P. orientalis trees that were at least 90 years of age were selected for sampling. Two or three cores were taken from each tree at breast height (1.3 m) from different sections of the stem using a 5-mm increment borer (a total of 92 cores were collected). To prevent contamination from other carbon sources, the collected samples were stored in glass tubes.
Ring-width, BAI, and stable carbon isotope analysis. The core samples were sanded with grain paper that varied from 100 to 800 mesh to make the tree-ring more clearly visible for cross-dating. Then, COFECHA was used to check dating, find errors in the individual series, and improve the cross-dating accuracy 7,28 . Eight of the cores had rotted or had been damaged by woodworms, and 11 of the cores were missing rings, which would cause inaccuracies in the test data. Cores with poor quality were excluded from analysis. Thus, only 71 cores were selected for analysis, and these were divided into two sections: 35 cores were used for width measurements, and the remaining 36 cores were used to evaluate stable carbon isotopes. After natural drying, fixation, and grinding, tree ring widths were measured at a resolution of 0.01 mm from each of the cores using LINTAB 6 measurement equipment (Frank Rinn, Heidelberg, Germany) and the data were analyzed using Time Series Analysis  and Presentation (TSAP) software package (Frank Rinn, Heidelberg, Germany). Cross-dating of the tree-ring data was verified using COFECHA, which assesses the quality of cross-dating and the measurement accuracy of tree-ring series using a segmented time-series correlation technique 28 . Tree-ring width decreases with age in mature trees due to the diameter growth of the stem. Therefore, BAI is more appropriate than DBHI (diameter at breast height increment) for modeling tree growth and forest productivity. BAI was estimated according to the following formula: where R is the radius at breast height and n is the year of tree-ring formation. We measured tree-ring width of 35 cores included 12 tree-ring cores more than 150 years older. We calculated BAI with the beginning width in 1855 calendar year, then we gave up the first 25 years data of each core considering the less number of sample cores and "juvenile age effects" might rise the errors of tree-ring widths (similar consideration applied to the analysis of tree-ring δ 13 C) 13 . Then we calculated the BAI between 1880 to 2014 of each tree.
Prior to the isotope analysis, the first 25 years of each core were removed to avoid "juvenile age effects" on the tree-ring isotopes 13 . In this study, pooling was necessary because P. orientalis is a slow-growing conifer with very narrow rings, making it difficult to obtain enough material to measure the individual isotopic series. In addition, the small ring size would have created an unacceptably high risk of measurement error. Several recent studies have tested and proved the representativeness of pooled isotopic series for tree-ring δ 13 C compared to individual isotopic series [29][30][31][32][33][34] . The cores were divided into individual rings using a scalpel to cut along ring lines under a stereomicroscope (40× magnification) and rings within the same year were pooled. Earlywood and latewood were not separated for the isotopic analyses, as recent studies have shown that there is no difference between the two wood types at an isotopic level 35,36 . Each pooled sample was ground into a powder using a ball mill, followed by sieving through an 80-mesh sieve. Cellulose was extracted from all samples to avoid isotope variations during the iWUE calculation, so that they were based purely on the changes in the relative abundance of individual wood constituents that typically have different isotope signatures 37 . The method for cellulose extraction was modified from Loader 38 . Cellulose samples and whole wood samples were used for δ 13 C analysis using a high temperature reactor (HT-O, HEKAtech GMBH, Wegberg, Germany) coupled to a DELTAplus XP Mass Spectrometer (ThermoFinnigan, Thermo Fisher Scientific Inc., Waltham, MA, USA). Each sample was analyzed three times, and the average values were calculated.
The results from the isotope ratio deviations are presented using the common δ notation: where R refers to the ratio of the 13 C to 12 C isotopes in the sample ('sa') and the reference ['re' , compared to the PDB (Pee Dee Belemnite) standard]. Therefore, cellulose Δ 13 C was calculated as follows: where δ 13 C a and δ 13 C p are the ratios in atmospheric CO 2 and tree-ring cellulose, respectively. The relative rates of carbon fixation and stomatal conductance are the primary factors that determine Δ. According to the model proposed by Farquhar 12 , Δ 13 C and δ 13 C have different trends.
where a is the discrimination against 13 CO 2 during CO 2 diffusion through the stomata (a = 4·4‰), b is the discrimination associated with carboxylation (b = 27‰), and Ci and Ca are the intercellular and ambient CO 2 concentrations, respectively. Fick's law is as follows: The ratio of leaf conductance to water vapor is 1.6 g CO 2 , and the change in 13 C can be related to the A g / H O 2 ratio as follows: iWUE was calculated using the following equation: Climate data, CO 2 records, and δ 13 C a . All weather data were obtained from the Miyun National Forest Farm Meteorological Station (117°07′E, 40°22′N) located at 352 m a.s.l., 5 km from the study area. The data were checked for quality and consistency, and included the annual minimum, mean, and maximum temperatures, as well as annual precipitation, from 1951 to 2014. As Fig. 1  Statistical analysis. To model tree growth, we measured all cores ring widths and calculated BAI of each tree and then averaged them. One-way Analysis of Variance (ANOVA) was used to examine the BAI variation among trees. We used pooled wood samples, mixing material of the same cross-dated ring for measuring annual δ 13 C and then calculated iWUE. Regression analyses and t-test were used to identify significant trends in δ 13 C, Δ 13 C, iWUE, and BAI from 1880 to 2014, and to test the relationships between iWUE and BAI, between iWUE and environmental factors (include atmospheric CO 2 concentration, annual temperature and precipitation). R 2 and slope p-values were also estimated. All statistical analyses were carried out using the SPSS 11.0 statistical software package (SPSS, Chicago, IL, USA).

Results
Patterns of δ 13 C, Δ 13 C, iWUE, and BAI. Figure 2 shows the whole wood δ 13 C w and cellulose δ 13 C c trends under decreased stable atmospheric carbon isotopes (δ 13 C a ) ("Suess effect") from 1880 to 2014. The t-test revealed that a significant difference between δ 13 C w and δ 13 C c , implying that δ 13 C c cannot be replaced by δ 13 C w . We detected a decreased trend in both δ 13 C w and δ 13 Cc curves. The δ 13 C c pattern could be divided into approximately three periods, i.e., two plateau periods from 1880 to 1917 and from 1975 to 2014 and one decreasing period from 1918  orientalis, the tree ring whole wood stable carbon isotope (δ 13 C w , circles) record for P. orientalis, and the atmospheric stable carbon isotope (δ 13 C a , dotted line) between 1880 and 2014. We connected the annual δ 13 Cc points to a polyline, for convenient comparison with δ 13 C w .  (Fig. 3), while the discrimination (Δ 13 C) trends exhibited the opposite pattern (a decrease). There was a significant increase in iWUE, at a rate of 0.49 μmol mol −1 per year, throughout the last century (p < 0.05). iWUE ranged from 82.35 μmol mol −1 (1920) to 138.12 μmol mol −1 (2014) during the study period, and a significant increase was detected from 1974 to 2014, at a rate of 1.1 μmol mol −1 per year. Higher Ci values were associated with lower discrimination (Δ 13 C) and higher iWUE.
Tree-ring width tended to fluctuate during the last century, ranging from 0.029 cm to 0.193 cm, with an average of 0.1 cm (Fig. 4). Tree-ring width increased initially (from 1880 to 1961) and then decreased from 1962 to 2001, as shown in Fig. 4. Since 2002, there has been a slight increase in tree-ring width. The averaged BAI showed a significant quadratic curve increase from 1880 to 2014 (R 2 = 0.6353, p < 0.01), especially during the early years of growth, after which the increasing trend weakened since 1960. The increase in BAI in the earlier decades ranged from 2.01 cm 2 to 19.7 cm 2 , which accounted for approximately 82.7% of the total increase, indicating that the trees grew rapidly during that period. The rapid increasing phase was followed by a decrease in BAI, which has fluctuated since the 1960s.

Relationships between iWUE and temperature, precipitation, and atmospheric CO 2 .
Correlations between iWUE and annual temperature were significant and positive (R 2 = 0.5537, p < 0.01, Fig. 5). The slope of the regression was 9.4669, indicating that every 0.1 °C increase in annual temperature was accompanied by an by a 1.21 μmol mol −1 increase in iWUE per year. Annual precipitation fluctuated greatly, ranging from 250 mm to 800 mm, and decreased in recent years. However, no significant relationship between iWUE and precipitation was observed. The results of the correlation analysis indicated that iWUE was positively correlated with elevated Ca (r 2 = 0.7997, p < 0.01).
Relationships between BAI and temperature, precipitation, Ca, and iWUE. No significant relationship was observed between BAI and annual temperature (R 2 = 0.0091, p > 0.05, Fig. 6), indicating that the  influence of annual temperature on BAI was not as strong as that on iWUE. In addition, the relationship between BAI and precipitation were weak and not significant (R 2 = 0.0048, p > 0.05). There was a nonlinear relationship between BAI and Ca. Elevated atmospheric CO 2 concentration significantly (p < 0.01) stimulated P. orientalis biomass accumulation when Ca was less than approximately 320 ppm in the early phase; however, this effect was not pronounced when Ca exceeded 320 ppm. The correlation between BAI and iWUE was positive and significant (p < 0.01, Fig. 7).

Discussion
We found that in the first half of the last century, the error in the P. orientalis tree-ring δ 13 C was relatively high, but it decreased gradually over time. It seems that saplings exhibited varying responses to environmental change; this, can be described as the "juvenile effect". One reason for this age-related trend in δ 13 C could be the recycling of respired air that has already been depleted of 13 C by young trees growing close to the forest floor. Plant tissue δ 13 C is significantly lower than that found in the atmosphere because atmospheric CO 2 contains 98.9% of light-stable carbon isotope and 1.1% heavy-stable carbon isotope, and the light-stable carbon isotope is preferentially absorbed during photosynthesis, resulting in carbon isotope fractionation, which is known as the "carbon isotope effect" 39 . This process may be influenced by biotic factors, such as tree age and height, as well as abiotic factors such as meteorological conditions (radiation regime, temperature, precipitation, atmospheric pressure, etc.), CO 2 concentration, site conditions (water availability, nutritional constraints, elevation and slope exposure, etc.), and forest management and anthropogenic perturbations (fires, climate change, etc.) 31,39-41 . Chen et al. 42 suggested that altitudinal variation in δ 13 C may be negative or unchanged for plants growing in arid and/ or semiarid regions, and atmospheric pressure has a negative effect on plant δ 13 C. Di Matteo et al. 43 conducted a replicated field trial to characterize the interacting effects of aspect and intensifying drought on foliar δ 13 C and to quantify temporal variation in stomatal conductance. They found a significant reduction in stomatal conductance (p < 0.01) as the drought increased along a sunny slope, and foliar δ 13 C increased significantly (p < 0.01). In this Figure 5. Relationships between iWUE and annual temperature, precipitation, and atmospheric CO 2 concentration (Ca). Lines indicate the change trends. The regression equations (iWUE and annual temperature, iWUE and annual precipitation, iWUE and annual atmospheric CO 2 concentration), coefficients, and levels of significance are also shown. study, we selected 10 typical plots, for which the altitude ranged from 332 m to 562 m and located on two slope aspects, to carry out our study with the original intention of understanding the average conditions of iWUE and tree growth of P. orientalis in northern China. Therefore, we did not take into account the effect of differences in exposure and elevation on tree-ring δ 13 C.
During the study period, the P. orientalis iWUE was significantly and positively correlated with annual temperature. This result agrees with research conducted on other tree species around the world 7,37,44 . This phenomenon can be explained by plant ecophysiology. First, when plants are subjected to high temperatures their stomata partially close, and then g s decreases to prevent an excessive loss of water, thereby reducing the internal CO 2 concentration of the leaves. The absorption and utilization rate of internal leaf CO 2 increase to maintain high photosynthetic rate (An). At the same time, an increase in temperature causes an increase in the vapor-pressure deficit and in the transpiration rate (Tr); accordingly, iWUE and temperature were positively related in this study. The number of days for which the daily maximum temperature was higher than 30 °C increased substantially in the study area each year, together with the frequency of extreme temperature values, leading to reduced stomatal conductance and altering the environmental and physiological functions of P. orientalis.
Following the second industrial revolution, atmospheric CO 2 concentrations have gradually increased, rising from 276.1 ppm to 397.6 ppm in northern China according to the China Global Atmosphere Watch Baseline Observatory. We found that the iWUE increased significantly under elevated atmospheric CO 2 concentrations. Our findings are consistent with other studies conducted on the effects of elevated CO 2 on leaf physiological activity, which demonstrating that increased CO 2 promotes WUE (determined by traditional measurement methods: weighing, gas exchange, etc.) 45,46 Jackson et al. 47 found that increased CO 2 led to 50% decreases in stomatal conductance and transpiration, while the photosynthetic rate increased by 70% and WUE nearly doubled with sufficient illumination. Curtis et al. 48 found that leaf photosynthetic products increased by approximately 50% under elevated CO 2 concentrations after combining 38 groups of experimental data. Gagen et al. 49 noted that northern boreal Pinus sylvestris may have reached a threshold with respect to its ability to increase WUE under rising atmospheric CO 2 concentrations.
In this study, we used the averaged BAI to characterize trees growth. Then we did the ANOVA between averaged BAI and each tree's BAI, a total of 35 pairs of data series, found that the 77.1% pairs had no significant difference. Thus the variation among trees could be ignored and averaged BAI could represent most trees growth in this region. The averaged BAI showed a significant increasing tendency (p < 0.01); however, the coefficient of determination between BAI and iWUE was relatively low (R 2 = 0.3299), and there was still some difference values between before and after the 1970s. The BAI increased significantly before the 1970s and then fluctuated but did not increase significantly. It seems that there is a limitation beyond which the increase of BAI would abate. Some scholars found that when atmospheric CO 2 concentration exceeds approximately 350 ppm, the BAI decreased with increasing atmospheric CO 2 concentration 4 . In this study, we found that the tipping point of the concentration for the BAI response curve was approximately 320 ppm. Thus, there is a contrast between the results of short-term experiments and long-term studies. A notable difference is nitrogen availability, which may positively stimulate leaf area, photosynthesis, and plant growth 50 . However, the effect of nitrogen on plants may be less pronounced or may even become negative over time due to altered soil N transformation processes and progressive nutrient limitations in ecosystems approaching and surpassing N saturation 4,51 . It is likely that these effects caused the difference in the results between short-term experiments and long-term studies.
In addition, there was no clear response of the iWUE to changes in precipitation. In the rocky mountainous area of northern China, precipitation mainly causes increases in runoff with some percolation through the soil. Changes in precipitation typically affect plant growth conditions by altering the soil moisture content 52 . There were no significant changes in precipitation during the study period and only a slight decrease was observed after the year 2000. In any case, the inter-annual variability of rainfall is very high, and we did not found any relationship between annual precipitation and iWUE or BAI. It seems that P. orientalis is well adapted to dry conditions and can cope withthe limitation posed by dry conditions by increasing iWUE. This finding agrees with that of Niinemets et al. 53 for evergreens that exhibited greater responsiveness to increased CO 2 . Many studies have focused on the relationship between tree growth and water use efficiency, but it is unclear whether increased water use efficiency translates into greater tree growth 54,55 . Sleen 56 found that while increasing atmospheric CO 2 concentrations generally result in more efficient water use in the monsoon forests of Thailand, tree growth still tends to decline. This means that changes in other limiting climatic factors exert stronger opposing effects on tree growth. Wang et al. 57 found that since 1800, tree radial growth was not always associated with increased iWUE in the Xinglong Mountains of northwestern China; in fact, tree growth rates declined during periods of drought and particularly in the past decade. These results suggested that increases in iWUE are unable to compensate for the negative effects of severe limitations on tree radial growth.
In contrast with other studies, we found evidence for long-term increases in BAI that were associated with increases in iWUE. P. orientalis growth increased as iWUE increased in the mountainous area of northern China due to drought tolerance and high temperature resistance. In addition, the long-term water use efficiency of the P. orientalis forest ecosystem increased under elevated CO 2 concentrations and rising temperatures, suggesting that trees overcame the unfavorable factors and made full use of favorable growth conditions. This is mainly because of the special leaf structure, i.e., photosynthetic limitation by mesophyll diffusion varies strongly with the leaf structure 58 . This limitation caused by internal diffusion conductance is expected to be particularly large in sclerophyllous evergreens such as P. orientalis, which have a structure that enables them to adapt to water limitations 59 . Furthermore, P. orientalis, supports numerous leaf age classes, leading to high leaf-area indices as 8-13 m 2 m −2 on a projected leaf area basis 60 . With a more robust leaf structure, typical reductions in stomatal conductance observed in response to elevated CO 2 are likely to alter photosynthesis reduction, and generate a stronger response of iWUE 53,61 . In the context of climate change, the eco-physiological functions of the P. orientalis forest ecosystem are not inhibited, and maintains a high potential for carbon sequestration.

Conclusions
Global climate change has had significant impacts on forest ecosystem function worldwide, and these effects vary across regions. Understanding the responses of iWUE and tree growth to climate change has great importance for predicting how plants may respond to climate variations in the future and its implications for forest management and conservation. In our study, we detected a significantly increasing trend in iWUE curves (p < 0.01) and a significant positive correlation between iWUE and BAI (p < 0.01). Long-term iWUE trend could be largely and positively driven by the elevated CO 2 concentration and temperature. Elevated atmospheric CO 2 concentration stimulated P. orientalis BAI accumulation significantly (p < 0.01) when Ca was below about 320 ppm in the early phase. Then the promotion effect were not pronounced when Ca was above 320 ppm. There was not a decrease in the growth of drought-tolerant conifers in northern China, suggesting that growth of P. orientalis in semi-arid areas benefited from the elevated atmospheric CO 2 concentrations. The study outcomes may be used to improve forest management and conservation in semi-arid mountainous areas.