- Email: [email protected]
Precipitation reconstruction for the southern Altay Mountains (China) from tree rings of Siberian spruce, reveals recent wetting trend
Dendrochronologia 32 (2014) 266–272
Contents lists available at ScienceDirect
Dendrochronologia journal homepage: www.elsevier.com/locate/dendro
ORIGINAL ARTICLE
Precipitation reconstruction for the southern Altay Mountains (China) from tree rings of Siberian spruce, reveals recent wetting trend Feng Chen a,b,c,∗ , Yu-jiang Yuan a,b , Wen-shou Wei a,b , Tong-wen Zhang a,b , Hua-ming Shang a,b , Ruibo Zhang a,b a Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China b Xinjiang Laboratory of Tree-ring Ecology, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China c MOE Key Laboratory of Western China’s Environmental Systems, Collaborative Innovation Centre for Arid Environments and Climate Change, Lanzhou University, Lanzhou 73000, China
a r t i c l e
i n f o
Article history: Received 12 August 2013 Accepted 16 June 2014 Keywords: Altay Mountains Picea obovata Ring width Climatic response Precipitation reconstruction
a b s t r a c t We developed six tree-ring width chronologies of Siberian spruce (Picea obovata) from the low elevation forest of the southern Altay Mountains in northern Xinjiang, China. Although the six chronologies come from different sampling sites, significant correlations existed among the chronologies (r ≥ 0.477), and the first principal component (PC1) accounted for 72.2% of total variance over their common period 1825–2010. Correlation response analysis revealed that radial growth of Siberian spruce is mainly limited by a 12-month precipitation starting from July of the previous year to June of the current year. We therefore developed a July–June precipitation reconstruction spanning 1825–2009, which explained 65.5% of the instrumental variance for the period 1962–2009. The information of our precipitation reconstruction suggested that dry conditions existed for the periods 1829–1838, 1852–1855, 1876–1888, 1898–1911, 1919–1923, 1932–1936, 1943–1955, 1963–1968, 1973–1984 and 2007–2009, and wet conditions for the periods AD 1825–1828, 1839–1851, 1856–1875, 1889–1897, 1912–1918, 1924–1931, 1937–1942, 1956–1962, 1969–1972 and 1985–2006. Spatial climate correlation analyses with gridded land surface data revealed that our precipitation reconstruction contains a strong precipitation signal for the Altay Mountain ranges. Our reconstruction agreed with the moisture-sensitive tree ring width series of Siberian larch from the Altay Mountains of Mongolia on a decadal timescale. In addition, in contrast to a drying trend in north central China, a clear wetting trend has occurred in the southern Altay Mountains since 1980s. © 2014 Elsevier GmbH. All rights reserved.
Introduction The Altay Mountains is a region with complex physical geography, climate interactions, high level of endemism in the flora and fauna, and a number of coniferous species preserved in natural ecosystems. Some of the most interesting habitats include the treeline forests. These are often dominated by centuries-old trees with high potential for the development of centennial-length tree-ring chronologies. Based on tree rings of Siberian larch (Larix sibirica), a number of dendroclimatic reconstructions for the Altay Mountains
∗ Corresponding author at: Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China. Tel.: +86 9912613017. E-mail address: [email protected] (F. Chen). http://dx.doi.org/10.1016/j.dendro.2014.06.003 1125-7865/© 2014 Elsevier GmbH. All rights reserved.
have been developed in recent decades (Ovtchinnikov et al., 2000; Frank et al., 2007; Myglan et al., 2008; Zhang et al., 2008; Davi et al., 2009; Sidorova et al., 2012; Chen et al., 2012). These studies have clearly demonstrated the potential for using Siberian larch to better understand past trend in climate, but no study has focused on the strength and clarity of the climatic response of Siberian spruce (Picea obovata) that dominate low elevation forests of the Altay Mountains. Some studies have caused much debate over climate change in the 20th century and especially about unprecedented warmth after the mid 1980s (Yuan et al., 2004; Shi et al., 2007; Chen et al., 2013a). The regional distribution of the increased extreme rain and attribution of precipitation variability in the arid Central Asia to specific climate forcing (e.g. global warming) are still uncertain, and increasing the confidence of future projection of rainfall pattern remains a challenge. The tree-ring width and density chronologies
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
267
Fig. 2. Mean temperature and precipitation at Fuyun, based on long-term averages (1962–2009).
Fig. 1. Location map of sampling sites and meteorological station.
provide a valuable tool for understanding of local climate change (Chen et al., 2013b). In this context it is of crucial importance to construct long and reliable tree-ring chronologies based on the different tree-ring parameters, particularly for regions where such records are scarce. In the present study, we developed tree-ring width chronologies of Siberian spruce trees grown in the low elevation area of the southern Altay Mountains (China). The results of climatic response are used to assess the usefulness of Siberian spruce for paleoclimate studies. Based on the network of tree-ring width chronologies, a reconstruction of annual precipitation spanning AD 1825–2009 is developed. We also compared the results of this study with those of the previous study (Davi et al., 2009; Chen et al., 2012) to assess the common climate signals of the Altay Mountains. Materials and methods Study area The study was conducted in Fuyun and Fuhai counties, located in the southern Altay Mountains, northern Xinjiang, China (Fig. 1). The study area is situated in the transitional zone from plateau in the east to plain in the west in terms of topography. Elevation in the study area ranges from 500 to 3000 m. The climate is characterized as temperate continental with short, cool summers and long, severe winters. The climatic conditions are influenced by arctic air masses, which come from the Arctic Ocean and move across the Ural Mountains without any topographic barrier far away into Siberia and the Altay Mountains. The mean annual precipitation is about 186.4 mm and the mean annual temperature is about 3.0 ◦ C. Snowfall usually lasts 6 months (from October to March). July is the hottest month (average temperature 22.2 ◦ C) while January is the coldest month (average temperature −20.5 ◦ C). This area is one of the coldest places in China during the winter. Permafrost is well represented in this area and seasonal thawing of soil ice does not exceed 50–200 cm in depth. In the study area, conifer forests above 2200 m in elevation are dominated by Siberian larch which usually forms single-species stands. Between the elevation of 1500 and 2200 m forests become dominated by Siberian spruce. Subalpine meadow occurs above 2700 m a.s.l. The cores of Siberian spruce were collected from six
low elevation sites (<1700 m). At these sites, as the indicators of drought stress, open-canopy spruce trees with sparse vegetation grow on thin or rocky soils. In combination, these six sites provide 315 samples taken from 169 trees (Table 1). Data collection and chronology development Increment cores were collected at breast height from dominant trees using an increment borer. We employed standard dendrochronological techniques in the processing of tree cores (Fritts, 1976; Cook and Kairiukstis, 1990). Ring widths were measured using a Velmex measuring system with accuracy of 0.001 mm. A thorough check was made of the strength of the cross matching between trees using the computer program COFECHA (Holmes, 1983). The data were then standardized based on the negative exponential function, using ARSTAN software (Cook, 1985), in order to remove the age-related growth trends and preserve the climate signal. The final site chronologies were computed by calculating bi-weighted robust means of annual tree-ring indices. We computed several statistical parameters commonly used in dendrochronology from the non-standardized tree-ring width series. The mean sensitivity (MS) measures year-to-year variation in treering width and is thus considered an estimate of the extent to which the chronology reflects local climate variation (Cook and Kairiukstis, 1990). The first-order autocorrelation (first AC) reflects the influence of previous year’s growth on current growth. The expressed population signal (EPS) quantifies the degree to which the constructed chronology portrays the hypothetically perfect one (Wigley et al., 1984). We used an EPS value of 0.85 as a threshold for the reliability of our chronologies. Meteorological data and statistical analysis Instrumental climate records of Fuyun (1962–2009) were obtained from the China National Climatic Data Center (Fig. 2). Bootstrapped correlation functions were computed using the software DENDROCLIM2002 to explore the climate-growth relationships (Biondi and Waikul, 2004). A 95% confidence level criterion was used to determine the statistical significance. The ring-width chronologies were compared with a 15-month window of climate data spanning the period from previous July through September of the current growing season. To assess the common growth forces among the individual sites of the southern Altay Mountains, correlation matrix and principal
268
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
Table 1 Site information for standardized tree-ring chronologies. Site TLD XSK SEE XTK KYS DEN
Lat. (N) ◦
47 49 47◦ 42 47◦ 35 47◦ 41 47◦ 31 47◦ 25
Long. (E) ◦
89 00 88◦ 59 88◦ 48 89◦ 06 89◦ 39 89◦ 38
Elevation (m)
Aspect
1260–1280 1130–1280 1155–1167 1667–1700 1590–1660 1430–1460
Slope
Core/tree number ◦
W EN NW S E E
30–40 5–40◦ 0–15◦ 30–45◦ 10–40◦ 10–40◦
57/29 46/26 48/25 51/28 51/27 62/34
Table 2 Summary of statistics for ring-width standard chronologies of Siberian spruce of the six sites in the southern Altay Mountains, China.
Chronology period The period with an EPS of at least 0.85 Mean correlation with master Mean sensitivity Standard deviation First-order autocorrelation Signal-to-noise ratio Variance in first eigenvector (%)
TLD
XSK
SEE
XTK
KYS
DEN
1799–2010 1825–2010 0.742 0.232 0.275 0.472 39.314 53.7
1687–2010 1715–2010 0.676 0.217 0.301 0.574 18.550 48.1
1668–2010 1721–2010 0.581 0.227 0.299 0.439 15.460 38.1
1638–2010 1691–2010 0.586 0.181 0.221 0.469 22.189 39.7
1783–2010 1815–2010 0.736 0.209 0.257 0.519 25.704 54.6
1762–2010 1788–2010 0.727 0.266 0.315 0.392 25.487 53.8
component analysis (PCA) was used in the intersite comparison over the 1825–2010 period, where the signal strength of the chronologies is acceptable based on the EPS criteria, as noted previously. A simple linear regression model was used to reconstruct the precipitation of the southern Altay Mountains. To demonstrate that our precipitation reconstruction represents regional precipitation variations, we conducted spatial correlations between our precipitation reconstruction and the updated 0.5 × 0.5 gridded July–June precipitation of CRU TS3.1 (Climate Research Unit) for the period 1963–2009 by the use of the KNMI climate explorer (http://climexp.knmi.nl). Results and discussion The characteristics of tree-ring chronologies The statistics of standard chronologies was given in Table 2. The six individual chronologies had an average correlation of the individual cores (with master series) of 0.675 (ranged from 0.581 to 0.742) and an average mean sensitivity (year to year variability) of 0.222 (ranged from 0.181 to 0.266). The high mean correlations with master series indicated good cross dating. The relatively high mean sensitivities and standard deviations (0.221–0.315) indicated rather immoderate interannual variations in the ring-width series. The variances in the first eigenvector (38.1–54.6%) indicated that the growth of different trees was responding to common factors. The first-order autocorrelation ranged from 0.392 to 0.574. This implied that the controlling factors (i.e., climate conditions, stand development and insect defoliators, etc.) that cause a ring to be narrow (or wide) in one year tend to carry over their effect on the growth of the following year. Our six chronologies ranged from 212 to 373 years in length (based on the oldest individual in the chronologies). The cross correlations in Table 3 showed highly significant similarities among the chronologies. Most of the correlation coefficients were significant at the 1% level. The highest correlation was found between KYS and DEN (r = 0.785, P < 0.01, n = 186). The PCA analyses conducted for the period 1825–2010 on the network of six standard chronologies showed that the first principal component (PC1) contains 72.2% of the total variance (Fig. 3). The strong common signals implied that the growth of spruce trees at the six sites expressed similar ecological characteristics and limited by large-scale common climate forcings.
Climate response analysis Fig. 4 shows the correlation coefficients for climate response analyses on monthly scale for six tree-ring chronologies from the southern Altay Mountains. The climate response analyses indicated that ring width at the all sites was positively correlated with July–August precipitation of the previous year and winter precipitation (especially December), and with May–July precipitation of the current growth year. The positive correlations with precipitation implied a dominant hydrological control on the growth of Siberian spruce. Above-average precipitation during late summer and early autumn might promote storage of carbohydrates and bud formation, thus enhancing growth during the following year (Fritts, 1974). Because the precipitation over the Altay Mountains in winter and early spring falls as snow, the positive correlations with winter precipitation may represent a relationship with snowpack and its subsequent effects on soil moisture (Chen et al., 2013a). Tree-growth benefits from the previous winter and current spring precipitation and snow, which increase the soil moisture content during the early phase of the growing season. Later on, with the rise in temperature, rapid expansion of tracheids and cell division in the cambium of trees at early stages of the growing season, the already existing water stress was accelerated in low elevation areas of the southern Altay Mountains. Increased May–July rainfall can help ease the threat of drought. As discussed above, Siberian spruce grows on thin or rocky soils (with low water-holding capacity) of the southern Altay Mountains might be affected by drought stress, and resulted in high sensitivity to precipitation changes. The strong association with August–September temperature was likely due to low temperature stress occurring during the late part of the growing season. Photosynthetic rates of plants were
Table 3 Cross correlations between the different ring width chronologies for the common time interval 1825–2010.
TLD XSK XTK KYS DEN SEE *
TLD
XSK
XTK
KYS
DEN
SEE
1.000 0.706* 0.641* 0.630* 0.641* 0.615*
1.000 0.722* 0.726* 0.719* 0.703*
1.000 0.731* 0.679* 0.634*
1.000 0.785* 0.558*
1.000 0.477*
1.000
Significant at the 1% level.
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
269
Fig. 3. Plot of the standard chronologies of Siberian spruce in the southern Altay Mountains, and the sample size and PC1.
generally temperature dependent, and therefore, low temperatures during the growing season reduced photosynthetic production for alpine and subalpine plants (Körner, 1999). Although autumn temperature (September) was still high (>10 ◦ C), the length of the growing season was shortened under low temperature condition, especially the low mean minimum temperature, and latewood development was also reduced under low temperature conditions (Li et al., 2013). July–June precipitation reconstruction To investigate the common growth forces, we screened PC1 in correlation analysis with the seasonal combinations of temperatures and precipitation from previous July to current September. The strongest correlation was found between the PC1 and the precipitation from previous July to current June (r = 0.809, n = 47, P < 0.01). Therefore, the reconstruction was performed by calibrating PC1 with total July–June precipitation data. As shown in Table 4, the calibration and verification results for the split subperiods, i.e., 1963–1985 and 1986–2009, generally showed a good model fit (Fig. 5A). The reduction error (RE) and the coefficient of efficiency (CE) based on the respective tests (P < 0.001) were positive and high, providing further evidence that the model has significant skill (Fritts, 1976). For the full calibration period (1963–2009),
the correlation between tree-ring width series and instrumental precipitation records is 0.809, which accounts for 65.5% of the precipitation variability from prior July to current June. By applying the regression model, the annual precipitation (previous July–current June) was recovered for the period AD 1825–2009 (Fig. 5B). The reconstructed precipitation ranged from 79.6 mm to 307.0 mm with a mean and a standard deviation () of 185.4 mm and 41.9 mm, respectively. Wet (dry) periods were defined as having an 10-year low-pass filter value that was continuously higher (lower) than the long-term average from AD 1825 to 2009. According to this definition, regional wet conditions occurred during AD 1825–1828, 1839–1851, 1856–1875, 1889–1897, 1912–1918, 1924–1931, 1937–1942, 1956–1962, 1969–1972 and 1985–2006; dry periods were identified during AD 1829–1838, 1852–1855, 1876–1888, 1898–1911, 1919–1923, 1932–1936, 1943–1955, 1963–1968, 1973–1984 and 2007–2009. The values beyond the inner horizontal lines (±1 SD) indicate dry and wet years, and those beyond the outer horizontal lines (±2 SD) indicate extremely dry and wet years. Seven extremely dry years (1885, 1900, 1920, 1945, 1951, 1974 and 1978) and 27 dry years were found in our precipitation reconstruction. The most severe drought in the past 185 years occurred in 1945. This severe drought has also been recorded in tree rings from Tien shan (Li et al., 2010). It is noteworthy that the extreme drought events in 1916–1919 and
270
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
Fig. 4. Climatic responses of tree-ring width chronologies of Siberian spruce in the southern Altay Mountains, obtained from simple correlations (bars). The dot lines indicate significant variables (P < 0.05).
1926–1932 (Esper et al., 2001; Liang et al., 2003; Chen et al., 2013a) were not significant in our reconstruction, implying that the Altay Mountains has its own unique climate characteristic.
species co-varied, indicating that precipitation was a strong common factor influencing tree growth in the low elevation forests of the Altay Mountains.
Regional-scale precipitation signals in the Altay Mountains
Tree growth and recent climate trends
The results of spatial correlation analysis showed that our precipitation reconstruction correlated >0.6 with July–June precipitation grid-box data in a large area of the Altay Mountains, with highest correlations occurring in the southern Altay Mountains (Fig. 5C). Despite a complex mountain terrain and spatial differences in local precipitation and tree growth, based on the above analysis results, July–June precipitation was found to be the most important force on the Siberian spruce growth of the southern Altay Mountains. Several moisture-sensitive tree ring width series of Siberian larch have been developed from the eastern side of the Altay Mountains, Mongolia (Davi et al., 2009). The Mongolia chronology has adequate replication after AD 1825, comparable with our precipitation reconstruction. Although the study areas are 200 km apart and use different species, high correlation coefficient (r = 0.25, P < 0.001) between our precipitation reconstruction and the Mongolia chronology was found, and increasing coherence after 10-year low-pass filtering (r = 0.45, P < 0.001). Both showed a very dry period in the 1880s and an increasing moisture trend in the late twentieth century (Fig. 6). The long-term growth trends of all two
The most noteworthy feature of our precipitation reconstruction occurred in the late 20th century. Along with the global warming of the late 20th century, the precipitation reconstruction for the southern Altay Mountains exhibits a upward trend from 1980 to 2001. On the basis of the available meteorological records, the study has evidenced a wetting trend in northwest China since the 1980s (Shi et al., 2007). Our reconstruction is in basic agreement with this assessment. This wetting trend is in contrast to a drying trend in north China, which has been attributed to the increased strength of the Westerlies. The increase in the strength of the Westerlies appears to be related to warming of sea surface temperatures over the North Atlantic and Indo–West Pacific Oceans, which is considered to be part of the global warming trend (Chen et al., 2013a). Unlike widely reported ‘divergence problem’ in northern forests (Briffa et al., 1998; D’Arrigo et al., 2008), larch and spruce trees in the Altay Mountains do not clearly lose their climate sensitivity under most recent warming (Chen et al., 2012 and this study), and the correlations of the tree growth with climate factors are still very high (r ≥ 0.7). The ring-width chronologies of Siberian spruce and Siberian larch from the Altay Mountains allow
Table 4 Calibration and verification statistics for the precipitation reconstruction.
r r2 RE CE Sign test
Calibration (1987–2009)
Verification (1963–1986)
Calibration (1963–1986)
0.696 0.484
0.822 0.676 0.671 0.617 20+/3−
0.822 0.676
Verification (1987–2009) 0.696 0.484 0.477 0.371 18+/6−
Full calibration (1963–2009) 0.809 0.655
38+/9−
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
271
Fig. 5. (A) Actual and reconstructed July–June precipitation during their common period 1963–2009. The estimation explains 65.5% of the actual precipitation variance in this common period. (B) Estimated (thin line) and 10-year low-pass filter (thick line) values of July–June precipitation. Central horizontal line shows the mean of the estimated values; inner dotted horizontal lines show the border of one standard deviation, and outer horizontal lines show the border of two standard deviations. (C) Spatial correlations between the precipitation reconstruction and the gridded precipitation data set (July–June). The analyses were performed using the KNMI climate explorer (http://climexp.knmi.nl).
Fig. 6. Comparison of our precipitation reconstruction against the moisture-sensitive tree ring width series in the eastern side of the Altay Mountains, Mongolia (Davi et al., 2009).
detecting the recent climate change (especially climate interactions) in a long-term context. Conclusions We have developed six tree-ring width chronologies of Siberian spruce from the low elevation forest of the southern Altay Mountains in northern Xinjiang, China. Very high cross correlations
were found among tree-ring width chronologies. High PC1 value (72.2%) means the strong common growth response to regional climatic variations, and PC1 can be used to evaluate the regional climate-growth relationships and to indicate regional climate variability. The climate response analysis shows that the ring-widths of Siberian spruce provide the best information for annual (July–June) precipitation reconstruction. While the climatic responses of ringwidth varied among the sites, ring-width series were strongly
272
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
correlated with August–September temperatures. According to the relationship between ring – width and precipitation, annual precipitation has been reconstructed from prior July to current June for the period from AD 1825 to 2009. This reconstruction accounts for 65.5% of the variance in instrumental data over the 1963–2009 period. Spatial correlation analysis and interspecific comparisons revealed that precipitation is the most important common forces on the tree growth of the low elevation forest of the Altay Mountains. The sensitivity of recent tree-growth to precipitation in the southern Altay Mountains was not reduced significantly under climate warming. As indicated in our study, regional precipitation variability has increased, and a clear wetting trend has occurred since the 1980s. Our results are preliminary and require confirmation from ongoing dendroclimatological studies of the Altay Mountains. Thus, further efforts should be taken to develop more comprehensive tree-ring networks, and to combine various tree-ring parameters like maximum latewood density and stable isotopes, to shed more light on the past climate variability of the Altay Mountains over long temporal and large spatial scales. Acknowledgments This work was supported by the National Science Foundation of China (No. 41275120), the Meteorology Public Welfare Industry Research Special Project (GYHY201206014), China Desert Meteorological Science Research Foundation (SQJ2013015). Tree-ring data from Mongolia were obtained from the NOAA Paleoclimatology International Tree-Ring Data Bank (ITRDB). Particular thanks are extended to the reviewers for their valuable suggestions and comments regarding the revision of the manuscript. References Biondi, F., Waikul, K., 2004. DENDROCLIM2002: a C++ program for statistical calibration of climate signals in tree-ring chronologies. Comput. Geosci. 30, 301–311. Briffa, K.R., Schweingruber, F.H., Jones, P.D., Osborn, T., 1998. Reduced sensitivity of recent tree growth to temperature at high northern latitudes. Nature 391, 678–682. Chen, F., Yuan, Y.J., Wei, W.S., Yu, S.L., Zhang, T.W., 2012. Climatic response of ring width and maximum latewood density of Larix sibirica in the Altay Mountains, reveals recent warming trends. Ann. Forest Sci. 69, 723–733. Chen, F., Yuan, Y.J., Chen, F.H., Wei, W.S., Yu, S.L., Chen, X.J., Fan, Z.A., Zhang, R.B., Zhang, T.W., Shang, H.M., Qin, L., 2013a. A 426-year drought history for Western Tian Shan, Central Asia inferred from tree-rings and its linkages to the North Atlantic and Indo–West Pacific Oceans. The Holocene 23, 1095–1104.
Chen, F., Yuan, Y.J., Wei, W.S., Zhang, R.B., Yu, S.L., Shang, H.M., Zhang, T.W., Qin, L., Wang, H.Q., Chen, F.H., 2013b. Tree-ring-based annual precipitation reconstruction for the Hexi Corridor, NW China: consequences for climate history on and beyond the mid-latitude Asian continent. Boreas 42, 1008–1021. Cook, E.R., 1985. A Time-Series Analysis Approach to Tree-Ring Standardisation. University of Arizona, Tucson. Cook, E., Kairiukstis, L., 1990. Methods of Dendrochronology: Applications in the Environmental Sciences. Kluwer Academic Publishers, Dordrecht, pp. 394pp. D’Arrigo, R., Wilson, R., Liepert, B., Cherubini, P., 2008. On the ‘divergence problem’ in northern forests: a review of the treering evidence and possible causes. Glob. Planet. Change 60, 289–305. Davi, N.K., Jacoby, G.C., D’Arrigo, R.D., Baatarbileg, N., Li, J., Curtis, A.E., 2009. A treering-based drought index reconstruction for far-western Mongolia: 1565–2004. Int. J. Climatol. 29, 1508–1514. Esper, J., Treydte, K., Gärtner, H., Neuwirth, B., 2001. A tree-ring reconstruction of climatic extreme years since AD 1427 for Western Central Asia. Palaeobotanist 50, 141–152. Frank, D.C., Ovchinnikov, D., Kirdyanov, A., Esper, J., 2007. The potential for longterm climatic reconstructions in the Central Altay Mountains from living and relict larch. Trace 5, 85–96. Fritts, H.C., 1974. Relationships of ring widths in arid-site conifers to variations in monthly temperature and precipitation. Ecol. Monogr. 44, 411–440. Fritts, H.C., 1976. Tree Rings and Climate. Academic Press, London, pp. 567pp. Holmes, R.L., 1983. Computer assisted quality control in tree-ring dating and measurement. Tree-Ring Bull. 44, 69–75. Körner, C.H., 1999. Alpine Plant Life. Springer, Berlin, pp. 338pp. Li, J., Cook, E.R., Chen, F., Gou, X.H., D’Arrigo, R., Yuan, Y., Li, J., 2010. An extreme drought event in the central Tien Shan area in the year 1945. J. Arid Environ. 74, 1225–1231. Li, X., Liang, E., Gricar, J., Prislan, P., Rossi, S., Cufar, K., 2013. Age-dependence of xylogenesis and its climatic sensitivity in Smith fir on the south-eastern Tibetan Plateau. Tree Physiol. 33, 48–56. Liang, E., Shao, X., Kong, Z., Lin, J., 2003. The extreme drought in the 1920s and its effect on tree growth deduced from tree-ring analysis, a case study in North China. Ann. Forest Sci. 60, 145–152. Myglan, V.S., Oidupaa, O.C., Kirdyanov, A.V., Vaganov, E.A., 2008. 1929-year treering chronology for the Altai-Sayan. Region (Western Tuva). Archaeol. Ethnol. Anthropol. Eurasia 36, 5–31. Ovtchinnikov, D., Adamenko, M., Panushkina, I.A., 2000. 1105-year tree-ring chronology in Altai Region and its application for reconstruction of summer temperatures. Geolines 11, 121–122. Shi, Y., Shen, Y., Kang, E., Li, D., Ding, Y., Zhang, G., Hu, R., 2007. Recent and future climate change in northwest China. Clim. Change 80, 379–393. Sidorova, O.V., Saurer, M., Myglan, V.S., Eichler, A., Schwikowski, M., Kirdyanov, V., Bryukhanova, V., Gerasimova, O.V., Kalugin, I.A., Daryin, A.V., Siegwolf, R.T.W., 2012. A multi-proxy approach for revealing recent climatic changes in the Russian Altai. Clim. Dyn. 38, 175–188. Wigley, T.M.L., Briffa, K.R., Jones, P.D., 1984. On the average of correlated time series, with applications in dendroclimatology and hydrometeorology. J. Clim. Appl. Meteorol. 23, 201–213. Yuan, Y.J., He, Q., Yu, S.L., 2004. Feature of annual precipitation change in Tianshan Mountain area for the recent 40 years and comparison with those in the south and north Xinjiang. Acta Meteorol. Sin. 24, 220–226. Zhang, T.W., Yuan, Y.J., Yu, S.L., Wei, W.S., Yang, Q., Shang, H.M., 2008. June to September precipitation series of 1481–2004 reconstructed from tree-ring in the Western region of Altay prefecture, Xinjiang. J. Glaciol. Geocryol. 30, 659–667.
Contents lists available at ScienceDirect
Dendrochronologia journal homepage: www.elsevier.com/locate/dendro
ORIGINAL ARTICLE
Precipitation reconstruction for the southern Altay Mountains (China) from tree rings of Siberian spruce, reveals recent wetting trend Feng Chen a,b,c,∗ , Yu-jiang Yuan a,b , Wen-shou Wei a,b , Tong-wen Zhang a,b , Hua-ming Shang a,b , Ruibo Zhang a,b a Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China b Xinjiang Laboratory of Tree-ring Ecology, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China c MOE Key Laboratory of Western China’s Environmental Systems, Collaborative Innovation Centre for Arid Environments and Climate Change, Lanzhou University, Lanzhou 73000, China
a r t i c l e
i n f o
Article history: Received 12 August 2013 Accepted 16 June 2014 Keywords: Altay Mountains Picea obovata Ring width Climatic response Precipitation reconstruction
a b s t r a c t We developed six tree-ring width chronologies of Siberian spruce (Picea obovata) from the low elevation forest of the southern Altay Mountains in northern Xinjiang, China. Although the six chronologies come from different sampling sites, significant correlations existed among the chronologies (r ≥ 0.477), and the first principal component (PC1) accounted for 72.2% of total variance over their common period 1825–2010. Correlation response analysis revealed that radial growth of Siberian spruce is mainly limited by a 12-month precipitation starting from July of the previous year to June of the current year. We therefore developed a July–June precipitation reconstruction spanning 1825–2009, which explained 65.5% of the instrumental variance for the period 1962–2009. The information of our precipitation reconstruction suggested that dry conditions existed for the periods 1829–1838, 1852–1855, 1876–1888, 1898–1911, 1919–1923, 1932–1936, 1943–1955, 1963–1968, 1973–1984 and 2007–2009, and wet conditions for the periods AD 1825–1828, 1839–1851, 1856–1875, 1889–1897, 1912–1918, 1924–1931, 1937–1942, 1956–1962, 1969–1972 and 1985–2006. Spatial climate correlation analyses with gridded land surface data revealed that our precipitation reconstruction contains a strong precipitation signal for the Altay Mountain ranges. Our reconstruction agreed with the moisture-sensitive tree ring width series of Siberian larch from the Altay Mountains of Mongolia on a decadal timescale. In addition, in contrast to a drying trend in north central China, a clear wetting trend has occurred in the southern Altay Mountains since 1980s. © 2014 Elsevier GmbH. All rights reserved.
Introduction The Altay Mountains is a region with complex physical geography, climate interactions, high level of endemism in the flora and fauna, and a number of coniferous species preserved in natural ecosystems. Some of the most interesting habitats include the treeline forests. These are often dominated by centuries-old trees with high potential for the development of centennial-length tree-ring chronologies. Based on tree rings of Siberian larch (Larix sibirica), a number of dendroclimatic reconstructions for the Altay Mountains
∗ Corresponding author at: Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China. Tel.: +86 9912613017. E-mail address: [email protected] (F. Chen). http://dx.doi.org/10.1016/j.dendro.2014.06.003 1125-7865/© 2014 Elsevier GmbH. All rights reserved.
have been developed in recent decades (Ovtchinnikov et al., 2000; Frank et al., 2007; Myglan et al., 2008; Zhang et al., 2008; Davi et al., 2009; Sidorova et al., 2012; Chen et al., 2012). These studies have clearly demonstrated the potential for using Siberian larch to better understand past trend in climate, but no study has focused on the strength and clarity of the climatic response of Siberian spruce (Picea obovata) that dominate low elevation forests of the Altay Mountains. Some studies have caused much debate over climate change in the 20th century and especially about unprecedented warmth after the mid 1980s (Yuan et al., 2004; Shi et al., 2007; Chen et al., 2013a). The regional distribution of the increased extreme rain and attribution of precipitation variability in the arid Central Asia to specific climate forcing (e.g. global warming) are still uncertain, and increasing the confidence of future projection of rainfall pattern remains a challenge. The tree-ring width and density chronologies
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
267
Fig. 2. Mean temperature and precipitation at Fuyun, based on long-term averages (1962–2009).
Fig. 1. Location map of sampling sites and meteorological station.
provide a valuable tool for understanding of local climate change (Chen et al., 2013b). In this context it is of crucial importance to construct long and reliable tree-ring chronologies based on the different tree-ring parameters, particularly for regions where such records are scarce. In the present study, we developed tree-ring width chronologies of Siberian spruce trees grown in the low elevation area of the southern Altay Mountains (China). The results of climatic response are used to assess the usefulness of Siberian spruce for paleoclimate studies. Based on the network of tree-ring width chronologies, a reconstruction of annual precipitation spanning AD 1825–2009 is developed. We also compared the results of this study with those of the previous study (Davi et al., 2009; Chen et al., 2012) to assess the common climate signals of the Altay Mountains. Materials and methods Study area The study was conducted in Fuyun and Fuhai counties, located in the southern Altay Mountains, northern Xinjiang, China (Fig. 1). The study area is situated in the transitional zone from plateau in the east to plain in the west in terms of topography. Elevation in the study area ranges from 500 to 3000 m. The climate is characterized as temperate continental with short, cool summers and long, severe winters. The climatic conditions are influenced by arctic air masses, which come from the Arctic Ocean and move across the Ural Mountains without any topographic barrier far away into Siberia and the Altay Mountains. The mean annual precipitation is about 186.4 mm and the mean annual temperature is about 3.0 ◦ C. Snowfall usually lasts 6 months (from October to March). July is the hottest month (average temperature 22.2 ◦ C) while January is the coldest month (average temperature −20.5 ◦ C). This area is one of the coldest places in China during the winter. Permafrost is well represented in this area and seasonal thawing of soil ice does not exceed 50–200 cm in depth. In the study area, conifer forests above 2200 m in elevation are dominated by Siberian larch which usually forms single-species stands. Between the elevation of 1500 and 2200 m forests become dominated by Siberian spruce. Subalpine meadow occurs above 2700 m a.s.l. The cores of Siberian spruce were collected from six
low elevation sites (<1700 m). At these sites, as the indicators of drought stress, open-canopy spruce trees with sparse vegetation grow on thin or rocky soils. In combination, these six sites provide 315 samples taken from 169 trees (Table 1). Data collection and chronology development Increment cores were collected at breast height from dominant trees using an increment borer. We employed standard dendrochronological techniques in the processing of tree cores (Fritts, 1976; Cook and Kairiukstis, 1990). Ring widths were measured using a Velmex measuring system with accuracy of 0.001 mm. A thorough check was made of the strength of the cross matching between trees using the computer program COFECHA (Holmes, 1983). The data were then standardized based on the negative exponential function, using ARSTAN software (Cook, 1985), in order to remove the age-related growth trends and preserve the climate signal. The final site chronologies were computed by calculating bi-weighted robust means of annual tree-ring indices. We computed several statistical parameters commonly used in dendrochronology from the non-standardized tree-ring width series. The mean sensitivity (MS) measures year-to-year variation in treering width and is thus considered an estimate of the extent to which the chronology reflects local climate variation (Cook and Kairiukstis, 1990). The first-order autocorrelation (first AC) reflects the influence of previous year’s growth on current growth. The expressed population signal (EPS) quantifies the degree to which the constructed chronology portrays the hypothetically perfect one (Wigley et al., 1984). We used an EPS value of 0.85 as a threshold for the reliability of our chronologies. Meteorological data and statistical analysis Instrumental climate records of Fuyun (1962–2009) were obtained from the China National Climatic Data Center (Fig. 2). Bootstrapped correlation functions were computed using the software DENDROCLIM2002 to explore the climate-growth relationships (Biondi and Waikul, 2004). A 95% confidence level criterion was used to determine the statistical significance. The ring-width chronologies were compared with a 15-month window of climate data spanning the period from previous July through September of the current growing season. To assess the common growth forces among the individual sites of the southern Altay Mountains, correlation matrix and principal
268
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
Table 1 Site information for standardized tree-ring chronologies. Site TLD XSK SEE XTK KYS DEN
Lat. (N) ◦
47 49 47◦ 42 47◦ 35 47◦ 41 47◦ 31 47◦ 25
Long. (E) ◦
89 00 88◦ 59 88◦ 48 89◦ 06 89◦ 39 89◦ 38
Elevation (m)
Aspect
1260–1280 1130–1280 1155–1167 1667–1700 1590–1660 1430–1460
Slope
Core/tree number ◦
W EN NW S E E
30–40 5–40◦ 0–15◦ 30–45◦ 10–40◦ 10–40◦
57/29 46/26 48/25 51/28 51/27 62/34
Table 2 Summary of statistics for ring-width standard chronologies of Siberian spruce of the six sites in the southern Altay Mountains, China.
Chronology period The period with an EPS of at least 0.85 Mean correlation with master Mean sensitivity Standard deviation First-order autocorrelation Signal-to-noise ratio Variance in first eigenvector (%)
TLD
XSK
SEE
XTK
KYS
DEN
1799–2010 1825–2010 0.742 0.232 0.275 0.472 39.314 53.7
1687–2010 1715–2010 0.676 0.217 0.301 0.574 18.550 48.1
1668–2010 1721–2010 0.581 0.227 0.299 0.439 15.460 38.1
1638–2010 1691–2010 0.586 0.181 0.221 0.469 22.189 39.7
1783–2010 1815–2010 0.736 0.209 0.257 0.519 25.704 54.6
1762–2010 1788–2010 0.727 0.266 0.315 0.392 25.487 53.8
component analysis (PCA) was used in the intersite comparison over the 1825–2010 period, where the signal strength of the chronologies is acceptable based on the EPS criteria, as noted previously. A simple linear regression model was used to reconstruct the precipitation of the southern Altay Mountains. To demonstrate that our precipitation reconstruction represents regional precipitation variations, we conducted spatial correlations between our precipitation reconstruction and the updated 0.5 × 0.5 gridded July–June precipitation of CRU TS3.1 (Climate Research Unit) for the period 1963–2009 by the use of the KNMI climate explorer (http://climexp.knmi.nl). Results and discussion The characteristics of tree-ring chronologies The statistics of standard chronologies was given in Table 2. The six individual chronologies had an average correlation of the individual cores (with master series) of 0.675 (ranged from 0.581 to 0.742) and an average mean sensitivity (year to year variability) of 0.222 (ranged from 0.181 to 0.266). The high mean correlations with master series indicated good cross dating. The relatively high mean sensitivities and standard deviations (0.221–0.315) indicated rather immoderate interannual variations in the ring-width series. The variances in the first eigenvector (38.1–54.6%) indicated that the growth of different trees was responding to common factors. The first-order autocorrelation ranged from 0.392 to 0.574. This implied that the controlling factors (i.e., climate conditions, stand development and insect defoliators, etc.) that cause a ring to be narrow (or wide) in one year tend to carry over their effect on the growth of the following year. Our six chronologies ranged from 212 to 373 years in length (based on the oldest individual in the chronologies). The cross correlations in Table 3 showed highly significant similarities among the chronologies. Most of the correlation coefficients were significant at the 1% level. The highest correlation was found between KYS and DEN (r = 0.785, P < 0.01, n = 186). The PCA analyses conducted for the period 1825–2010 on the network of six standard chronologies showed that the first principal component (PC1) contains 72.2% of the total variance (Fig. 3). The strong common signals implied that the growth of spruce trees at the six sites expressed similar ecological characteristics and limited by large-scale common climate forcings.
Climate response analysis Fig. 4 shows the correlation coefficients for climate response analyses on monthly scale for six tree-ring chronologies from the southern Altay Mountains. The climate response analyses indicated that ring width at the all sites was positively correlated with July–August precipitation of the previous year and winter precipitation (especially December), and with May–July precipitation of the current growth year. The positive correlations with precipitation implied a dominant hydrological control on the growth of Siberian spruce. Above-average precipitation during late summer and early autumn might promote storage of carbohydrates and bud formation, thus enhancing growth during the following year (Fritts, 1974). Because the precipitation over the Altay Mountains in winter and early spring falls as snow, the positive correlations with winter precipitation may represent a relationship with snowpack and its subsequent effects on soil moisture (Chen et al., 2013a). Tree-growth benefits from the previous winter and current spring precipitation and snow, which increase the soil moisture content during the early phase of the growing season. Later on, with the rise in temperature, rapid expansion of tracheids and cell division in the cambium of trees at early stages of the growing season, the already existing water stress was accelerated in low elevation areas of the southern Altay Mountains. Increased May–July rainfall can help ease the threat of drought. As discussed above, Siberian spruce grows on thin or rocky soils (with low water-holding capacity) of the southern Altay Mountains might be affected by drought stress, and resulted in high sensitivity to precipitation changes. The strong association with August–September temperature was likely due to low temperature stress occurring during the late part of the growing season. Photosynthetic rates of plants were
Table 3 Cross correlations between the different ring width chronologies for the common time interval 1825–2010.
TLD XSK XTK KYS DEN SEE *
TLD
XSK
XTK
KYS
DEN
SEE
1.000 0.706* 0.641* 0.630* 0.641* 0.615*
1.000 0.722* 0.726* 0.719* 0.703*
1.000 0.731* 0.679* 0.634*
1.000 0.785* 0.558*
1.000 0.477*
1.000
Significant at the 1% level.
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
269
Fig. 3. Plot of the standard chronologies of Siberian spruce in the southern Altay Mountains, and the sample size and PC1.
generally temperature dependent, and therefore, low temperatures during the growing season reduced photosynthetic production for alpine and subalpine plants (Körner, 1999). Although autumn temperature (September) was still high (>10 ◦ C), the length of the growing season was shortened under low temperature condition, especially the low mean minimum temperature, and latewood development was also reduced under low temperature conditions (Li et al., 2013). July–June precipitation reconstruction To investigate the common growth forces, we screened PC1 in correlation analysis with the seasonal combinations of temperatures and precipitation from previous July to current September. The strongest correlation was found between the PC1 and the precipitation from previous July to current June (r = 0.809, n = 47, P < 0.01). Therefore, the reconstruction was performed by calibrating PC1 with total July–June precipitation data. As shown in Table 4, the calibration and verification results for the split subperiods, i.e., 1963–1985 and 1986–2009, generally showed a good model fit (Fig. 5A). The reduction error (RE) and the coefficient of efficiency (CE) based on the respective tests (P < 0.001) were positive and high, providing further evidence that the model has significant skill (Fritts, 1976). For the full calibration period (1963–2009),
the correlation between tree-ring width series and instrumental precipitation records is 0.809, which accounts for 65.5% of the precipitation variability from prior July to current June. By applying the regression model, the annual precipitation (previous July–current June) was recovered for the period AD 1825–2009 (Fig. 5B). The reconstructed precipitation ranged from 79.6 mm to 307.0 mm with a mean and a standard deviation () of 185.4 mm and 41.9 mm, respectively. Wet (dry) periods were defined as having an 10-year low-pass filter value that was continuously higher (lower) than the long-term average from AD 1825 to 2009. According to this definition, regional wet conditions occurred during AD 1825–1828, 1839–1851, 1856–1875, 1889–1897, 1912–1918, 1924–1931, 1937–1942, 1956–1962, 1969–1972 and 1985–2006; dry periods were identified during AD 1829–1838, 1852–1855, 1876–1888, 1898–1911, 1919–1923, 1932–1936, 1943–1955, 1963–1968, 1973–1984 and 2007–2009. The values beyond the inner horizontal lines (±1 SD) indicate dry and wet years, and those beyond the outer horizontal lines (±2 SD) indicate extremely dry and wet years. Seven extremely dry years (1885, 1900, 1920, 1945, 1951, 1974 and 1978) and 27 dry years were found in our precipitation reconstruction. The most severe drought in the past 185 years occurred in 1945. This severe drought has also been recorded in tree rings from Tien shan (Li et al., 2010). It is noteworthy that the extreme drought events in 1916–1919 and
270
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
Fig. 4. Climatic responses of tree-ring width chronologies of Siberian spruce in the southern Altay Mountains, obtained from simple correlations (bars). The dot lines indicate significant variables (P < 0.05).
1926–1932 (Esper et al., 2001; Liang et al., 2003; Chen et al., 2013a) were not significant in our reconstruction, implying that the Altay Mountains has its own unique climate characteristic.
species co-varied, indicating that precipitation was a strong common factor influencing tree growth in the low elevation forests of the Altay Mountains.
Regional-scale precipitation signals in the Altay Mountains
Tree growth and recent climate trends
The results of spatial correlation analysis showed that our precipitation reconstruction correlated >0.6 with July–June precipitation grid-box data in a large area of the Altay Mountains, with highest correlations occurring in the southern Altay Mountains (Fig. 5C). Despite a complex mountain terrain and spatial differences in local precipitation and tree growth, based on the above analysis results, July–June precipitation was found to be the most important force on the Siberian spruce growth of the southern Altay Mountains. Several moisture-sensitive tree ring width series of Siberian larch have been developed from the eastern side of the Altay Mountains, Mongolia (Davi et al., 2009). The Mongolia chronology has adequate replication after AD 1825, comparable with our precipitation reconstruction. Although the study areas are 200 km apart and use different species, high correlation coefficient (r = 0.25, P < 0.001) between our precipitation reconstruction and the Mongolia chronology was found, and increasing coherence after 10-year low-pass filtering (r = 0.45, P < 0.001). Both showed a very dry period in the 1880s and an increasing moisture trend in the late twentieth century (Fig. 6). The long-term growth trends of all two
The most noteworthy feature of our precipitation reconstruction occurred in the late 20th century. Along with the global warming of the late 20th century, the precipitation reconstruction for the southern Altay Mountains exhibits a upward trend from 1980 to 2001. On the basis of the available meteorological records, the study has evidenced a wetting trend in northwest China since the 1980s (Shi et al., 2007). Our reconstruction is in basic agreement with this assessment. This wetting trend is in contrast to a drying trend in north China, which has been attributed to the increased strength of the Westerlies. The increase in the strength of the Westerlies appears to be related to warming of sea surface temperatures over the North Atlantic and Indo–West Pacific Oceans, which is considered to be part of the global warming trend (Chen et al., 2013a). Unlike widely reported ‘divergence problem’ in northern forests (Briffa et al., 1998; D’Arrigo et al., 2008), larch and spruce trees in the Altay Mountains do not clearly lose their climate sensitivity under most recent warming (Chen et al., 2012 and this study), and the correlations of the tree growth with climate factors are still very high (r ≥ 0.7). The ring-width chronologies of Siberian spruce and Siberian larch from the Altay Mountains allow
Table 4 Calibration and verification statistics for the precipitation reconstruction.
r r2 RE CE Sign test
Calibration (1987–2009)
Verification (1963–1986)
Calibration (1963–1986)
0.696 0.484
0.822 0.676 0.671 0.617 20+/3−
0.822 0.676
Verification (1987–2009) 0.696 0.484 0.477 0.371 18+/6−
Full calibration (1963–2009) 0.809 0.655
38+/9−
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
271
Fig. 5. (A) Actual and reconstructed July–June precipitation during their common period 1963–2009. The estimation explains 65.5% of the actual precipitation variance in this common period. (B) Estimated (thin line) and 10-year low-pass filter (thick line) values of July–June precipitation. Central horizontal line shows the mean of the estimated values; inner dotted horizontal lines show the border of one standard deviation, and outer horizontal lines show the border of two standard deviations. (C) Spatial correlations between the precipitation reconstruction and the gridded precipitation data set (July–June). The analyses were performed using the KNMI climate explorer (http://climexp.knmi.nl).
Fig. 6. Comparison of our precipitation reconstruction against the moisture-sensitive tree ring width series in the eastern side of the Altay Mountains, Mongolia (Davi et al., 2009).
detecting the recent climate change (especially climate interactions) in a long-term context. Conclusions We have developed six tree-ring width chronologies of Siberian spruce from the low elevation forest of the southern Altay Mountains in northern Xinjiang, China. Very high cross correlations
were found among tree-ring width chronologies. High PC1 value (72.2%) means the strong common growth response to regional climatic variations, and PC1 can be used to evaluate the regional climate-growth relationships and to indicate regional climate variability. The climate response analysis shows that the ring-widths of Siberian spruce provide the best information for annual (July–June) precipitation reconstruction. While the climatic responses of ringwidth varied among the sites, ring-width series were strongly
272
F. Chen et al. / Dendrochronologia 32 (2014) 266–272
correlated with August–September temperatures. According to the relationship between ring – width and precipitation, annual precipitation has been reconstructed from prior July to current June for the period from AD 1825 to 2009. This reconstruction accounts for 65.5% of the variance in instrumental data over the 1963–2009 period. Spatial correlation analysis and interspecific comparisons revealed that precipitation is the most important common forces on the tree growth of the low elevation forest of the Altay Mountains. The sensitivity of recent tree-growth to precipitation in the southern Altay Mountains was not reduced significantly under climate warming. As indicated in our study, regional precipitation variability has increased, and a clear wetting trend has occurred since the 1980s. Our results are preliminary and require confirmation from ongoing dendroclimatological studies of the Altay Mountains. Thus, further efforts should be taken to develop more comprehensive tree-ring networks, and to combine various tree-ring parameters like maximum latewood density and stable isotopes, to shed more light on the past climate variability of the Altay Mountains over long temporal and large spatial scales. Acknowledgments This work was supported by the National Science Foundation of China (No. 41275120), the Meteorology Public Welfare Industry Research Special Project (GYHY201206014), China Desert Meteorological Science Research Foundation (SQJ2013015). Tree-ring data from Mongolia were obtained from the NOAA Paleoclimatology International Tree-Ring Data Bank (ITRDB). Particular thanks are extended to the reviewers for their valuable suggestions and comments regarding the revision of the manuscript. References Biondi, F., Waikul, K., 2004. DENDROCLIM2002: a C++ program for statistical calibration of climate signals in tree-ring chronologies. Comput. Geosci. 30, 301–311. Briffa, K.R., Schweingruber, F.H., Jones, P.D., Osborn, T., 1998. Reduced sensitivity of recent tree growth to temperature at high northern latitudes. Nature 391, 678–682. Chen, F., Yuan, Y.J., Wei, W.S., Yu, S.L., Zhang, T.W., 2012. Climatic response of ring width and maximum latewood density of Larix sibirica in the Altay Mountains, reveals recent warming trends. Ann. Forest Sci. 69, 723–733. Chen, F., Yuan, Y.J., Chen, F.H., Wei, W.S., Yu, S.L., Chen, X.J., Fan, Z.A., Zhang, R.B., Zhang, T.W., Shang, H.M., Qin, L., 2013a. A 426-year drought history for Western Tian Shan, Central Asia inferred from tree-rings and its linkages to the North Atlantic and Indo–West Pacific Oceans. The Holocene 23, 1095–1104.
Chen, F., Yuan, Y.J., Wei, W.S., Zhang, R.B., Yu, S.L., Shang, H.M., Zhang, T.W., Qin, L., Wang, H.Q., Chen, F.H., 2013b. Tree-ring-based annual precipitation reconstruction for the Hexi Corridor, NW China: consequences for climate history on and beyond the mid-latitude Asian continent. Boreas 42, 1008–1021. Cook, E.R., 1985. A Time-Series Analysis Approach to Tree-Ring Standardisation. University of Arizona, Tucson. Cook, E., Kairiukstis, L., 1990. Methods of Dendrochronology: Applications in the Environmental Sciences. Kluwer Academic Publishers, Dordrecht, pp. 394pp. D’Arrigo, R., Wilson, R., Liepert, B., Cherubini, P., 2008. On the ‘divergence problem’ in northern forests: a review of the treering evidence and possible causes. Glob. Planet. Change 60, 289–305. Davi, N.K., Jacoby, G.C., D’Arrigo, R.D., Baatarbileg, N., Li, J., Curtis, A.E., 2009. A treering-based drought index reconstruction for far-western Mongolia: 1565–2004. Int. J. Climatol. 29, 1508–1514. Esper, J., Treydte, K., Gärtner, H., Neuwirth, B., 2001. A tree-ring reconstruction of climatic extreme years since AD 1427 for Western Central Asia. Palaeobotanist 50, 141–152. Frank, D.C., Ovchinnikov, D., Kirdyanov, A., Esper, J., 2007. The potential for longterm climatic reconstructions in the Central Altay Mountains from living and relict larch. Trace 5, 85–96. Fritts, H.C., 1974. Relationships of ring widths in arid-site conifers to variations in monthly temperature and precipitation. Ecol. Monogr. 44, 411–440. Fritts, H.C., 1976. Tree Rings and Climate. Academic Press, London, pp. 567pp. Holmes, R.L., 1983. Computer assisted quality control in tree-ring dating and measurement. Tree-Ring Bull. 44, 69–75. Körner, C.H., 1999. Alpine Plant Life. Springer, Berlin, pp. 338pp. Li, J., Cook, E.R., Chen, F., Gou, X.H., D’Arrigo, R., Yuan, Y., Li, J., 2010. An extreme drought event in the central Tien Shan area in the year 1945. J. Arid Environ. 74, 1225–1231. Li, X., Liang, E., Gricar, J., Prislan, P., Rossi, S., Cufar, K., 2013. Age-dependence of xylogenesis and its climatic sensitivity in Smith fir on the south-eastern Tibetan Plateau. Tree Physiol. 33, 48–56. Liang, E., Shao, X., Kong, Z., Lin, J., 2003. The extreme drought in the 1920s and its effect on tree growth deduced from tree-ring analysis, a case study in North China. Ann. Forest Sci. 60, 145–152. Myglan, V.S., Oidupaa, O.C., Kirdyanov, A.V., Vaganov, E.A., 2008. 1929-year treering chronology for the Altai-Sayan. Region (Western Tuva). Archaeol. Ethnol. Anthropol. Eurasia 36, 5–31. Ovtchinnikov, D., Adamenko, M., Panushkina, I.A., 2000. 1105-year tree-ring chronology in Altai Region and its application for reconstruction of summer temperatures. Geolines 11, 121–122. Shi, Y., Shen, Y., Kang, E., Li, D., Ding, Y., Zhang, G., Hu, R., 2007. Recent and future climate change in northwest China. Clim. Change 80, 379–393. Sidorova, O.V., Saurer, M., Myglan, V.S., Eichler, A., Schwikowski, M., Kirdyanov, V., Bryukhanova, V., Gerasimova, O.V., Kalugin, I.A., Daryin, A.V., Siegwolf, R.T.W., 2012. A multi-proxy approach for revealing recent climatic changes in the Russian Altai. Clim. Dyn. 38, 175–188. Wigley, T.M.L., Briffa, K.R., Jones, P.D., 1984. On the average of correlated time series, with applications in dendroclimatology and hydrometeorology. J. Clim. Appl. Meteorol. 23, 201–213. Yuan, Y.J., He, Q., Yu, S.L., 2004. Feature of annual precipitation change in Tianshan Mountain area for the recent 40 years and comparison with those in the south and north Xinjiang. Acta Meteorol. Sin. 24, 220–226. Zhang, T.W., Yuan, Y.J., Yu, S.L., Wei, W.S., Yang, Q., Shang, H.M., 2008. June to September precipitation series of 1481–2004 reconstructed from tree-ring in the Western region of Altay prefecture, Xinjiang. J. Glaciol. Geocryol. 30, 659–667.