|Home | About | Journals | Submit | Contact Us | Français|
Shrubs and dwarf shrubs are wider spread on the Tibetan Plateau than trees and hence offer a unique opportunity to expand the present dendrochronological network into extreme environments beyond the survival limit of trees. Alpine shrublands on the Tibetan Plateau are characterized by rhododendron species. The dendrochronological potential of one alpine rhododendron species and its growth response to the extreme environment on the south-east Tibetan Plateau were investigated.
Twenty stem discs of the alpine snowy rhododendron (Rhododendron nivale) were collected close to the tongue of the Zuoqiupu Glacier in south-east Tibet, China. The skeleton plot technique was used for inter-comparison between samples to detect the growth pattern of each stem section. The ring-width chronology was developed by fitting a negative exponential function or a straight line of any slope. Bootstrapping correlations were calculated between the standard chronology and monthly climate data.
The wood of snowy rhododendron is diffuse-porous with evenly distributed small-diameter vessels. It has well-defined growth rings. Most stem sections can be visually and statistically cross-dated. The resulting 75-year-long standard ring-width chronology is highly correlated with a timberline fir chronology about 200 km apart, providing a high degree of confidence in the cross-dating. The climate/growth association of alpine snowy rhododendron and of this timberline fir is similar, reflecting an impact of monthly mean minimum temperatures in November of the previous year and in July during the year of ring formation.
The alpine snowy rhododendron offers new research directions to investigate the environmental history of the Tibetan Plateau in those regions where up to now there was no chance of applying dendrochronology.
In recent years, interest has increased in dendrochronological studies of shrubs and dwarf shrubs. Despite difficulties in cross-dating, dendrochronological techniques have successfully been applied to shrubs in various parts of the world (e.g. Ferguson, 1959, 1964; Woodcook and Bradley, 1994; Schweingruber and Dietz, 2001; Cherubini et al., 2003; Verheyden et al., 2004; Rayback and Henry, 2005; Zalatan and Gajewski, 2006; Schmidt et al., 2006; Au and Tardif, 2007; Bär et al., 2007; Xiao et al., 2007; Sass-Klaassen et al., 2008). These studies showed that shrub species have a great potential to extend the existing dendrochronological networks into extreme environments beyond the survival limits of trees.
More and more dendrochronological studies based on coniferous tree species have been conducted on the Tibetan Plateau in the last decade (see a brief summary by Liang et al., 2008, 2009); however little is known as to whether shrub species in that area can be used for dendrochronological studies (Xiao et al., 2007). Shrubs and dwarf shrubs are widespread on the Tibetan Plateau (Chang, 1981) and therefore offer a unique opportunity to expand the present dendrochronological network higher up, far beyond the altitudinal tree limits into the alpine ecosystem.
South-east Tibet is characterized by a cold and humid climate and a high diversity of forest types with the highest timberlines in the world (Miehe et al., 2007). Alpine shrublands, characterized by rhododendrons of many species, predominate in the ecotone between the treeline and the alpine meadows (Chang et al., 1981). In addition, rhododendrons often grow precariously on cliffs and ledges. This is just the species to enable us to extend tree-ring-based climatic proxy records beyond the treeline in south-east Tibet. In spite of such an exciting perspective, its potential as an ecological indicator for global climate change and dendrochronological studies is still unknown. This gap in our knowledge largely hinders our efforts to understand how alpine rhododendron ecosystems respond to climate variability and how they will respond to future climate scenarios.
The objectives of the present study, therefore, are to describe the wood anatomy of the alpine snowy rhododendron, Rhododendron nivale, and investigate its dendrochronological potential and growth response to the extreme environment on the south-eastern Tibetan Plateau.
The study area is located in the Ranwu region on the northern slope of the Kangri Karpo Mts in south-east Tibet, China (Fig. 1). The south-west Asian monsoon reaches this region along the Yarlung Zangbo River, thus bringing abundant rainfall. As a result, numerous maritime glaciers developed around the mountains, among them the Zuoqiupu Glacier, one of the biggest maritime glaciers in south-east Tibet.
At the meteorological station in Bomi (28°43′N, 99°17′E, 2423 m a.s.l; Fig. 2), the long-term mean of annual precipitation is 836 mm and the annual mean temperature 8·7 °C, with July (mean temperature of 16·6 °C) and January (0·2 °C) the warmest and the coldest month, respectively. Around 51 % of annual precipitation falls between June and September. The Zuoqiupu Glacier has a mean annual net snow accumulation of around 3·5 m (B. Q. Xu, Institute of Tibetan Plateau Research, CAS, China, pers. comm.). The instrumental climate data, recorded since 1961, show a significantly increasing trend in mean, maximum and minimum temperatures in all seasons and in annual precipitation (Liang et al., 2009).
Close to the tongue of the Zuoqiupu Glacier, a patch of a snowy rhododendron (Rhododendron nivale) shrub forest covers a south-facing slope (29°14′N, 96°52′E) from 4250 to 4500 m a.s.l. The substrate consists of small- to medium-sized gravel with little organic material. Snowy rhododendron has a distinct main stem clearly differentiated from the branches. Its height varies from 60 to 90 cm. The trunks can reach diameters of up to 8 cm. At the lower elevation, snowy rhododendron forests are severely disturbed by human activities, mainly from cutting wood for fuel. Clearly, the snowy rhododendron forest is threatened in the study area.
In order to avoid human disturbances and to maximize the climatic signal, high-elevation sites (4450–4500 m a.s.l.) were targeted for sampling in October 2007. There, the individual shrubs come into contact to each other, thus forming a continuous cover of the ground. Clearly, this forest remained less disturbed. The study site represents the upper limit of the altitudinal distribution of snowy rhododendron in the study area. Normally, snowy rhododendrons have multiple and twisted stems. In this study, an effort was made to select 20 dominant snowy rhododendron individuals with single, upright stems of maximal diameters, allowing their maximum ages to be assessed. Because of the small-diameter stems (<8 cm) and high wood density, it is impossible to take increment cores. Instead, one stem section was cut close to the root collar, enabling the oldest part of an individual to be explored. As described by Kolishchuk (1990), an additional three or four cross-sections were taken at 15-cm intervals along the entire stems of two individuals to analyse intra-plant growth variability and to detect locally absent rings. Because of the desire to minimize detrimental impacts on such an ecologically important population, an extensive sample collection was not made.
The samples were processed following standard dendrochronological practices (Cook and Kairiukstis, 1990). It was found that snowy rhododendron produces distinct annual rings and has no rotten centre. Skeleton plots, according to Stokes and Smiley (1968), representing graphically the year-to-year ring-width variability were used for inter-comparison between samples to detect the growth pattern of each stem section and to overcome cross-dating problems due to discontinuous growth rings. After a rigorous cross-dating of the stem sections, ring widths were measured at a resolution of 0·01 mm, and then the quality of the cross-dating was checked using the COFECHA program (Holmes, 1983). It was confirmed that cambial activity started synchronously in all parts of two individuals based on the serial sectioning test. A few locally absent rings could be detected and dated. Eighteen out of the 20 stem sections were visually and statistically cross-dated. It seems to be easier to cross-date snowy rhododendron stems than those of dwarf shrubs (Woodcook and Bradely, 1994; Rayback and Henry, 2005; Zalatan and Gajewski, 2006; Au and Tardif, 2007; Bär et al, 2007).
The 18 cross-dated ring-width series were standardized using the ARSTAN program to remove the biological growth trend as well as other low-frequency variations due to stand dynamics (Cook and Kairiukstis, 1990). Apart from the wider innermost rings, there is no distinct age trend visible. The chronology was developed by fitting a negative exponential function or a straight line of any slope. This conservative detrending was used to preserve as much low-frequency variability as possible in the short tree-ring series while maximizing the climate signal. To reduce the influence of outliers in the assembly of the mean chronology, all detrended series were averaged to form one standard chronology using the bi-weight robust mean.
Several descriptive statistics, including mean sensitivity, mean series intercorrelation (RBAR) and expressed population signal (EPS; Briffa and Jones, 1990), were calculated for the standard chronology to allow comparisons with other chronologies. RBAR is the mean correlation coefficient among tree-ring series. The EPS assesses the degree to which the chronology represents a hypothetical chronology based on an infinite number of cores; an EPS ≥0·85 is often taken to identify the reliable part of a tree-ring chronology (Briffa and Jones, 1990).
To determine the climate–tree growth relationships, bootstrapping correlations (Biondi and Waikul, 2004) were calculated between the standard chronology and monthly climate data obtained from the Bomi meteorological station for the period 1961–2007. The climate variables include total monthly precipitation, the monthly mean temperature as well as monthly mean maximum and minimum temperature for a 15-month period from July of the year prior to ring formation to September of the year of ring formation.
The wood of snowy rhododendron is diffuse-porous with evenly distributed small-diameter vessels (radial diameter up to 35 µm) and multiseriate rays (Fig. 3). Since only upright stems were selected, the tree-ring widths were rather uniform around the circumference. The distinctly visible growth rings are annual due to a strong climatic seasonality in this alpine ecotone. They can be discriminated by a band of radially flattened latewood fibres along the ring boundary. The high vessel density (number of vessels mm−2) of snowy rhododendron, as well as of Nepalese rhododendron species reported by Noshiro and Suzuki (2001), may be due to growth stress under the harsh alpine environments. No false rings were identified in alpine snowy rhododendron at the study site which is at a much higher elevation than the study sites in Nepal. In contrast, many false rings were found in several Nepalese rhododendron species whose growth was supposed to suffer under drought stress (Noshiro and Suzuki, 2001). Other studies (e.g. Sass-Klaassen et al., 2008; Wimmer et al., 2000) also suggested that false-ring formation may be induced by water stress. The snowy rhododendron's growth at the study site is not likely to be limited by moisture availability.
The individuals selected were on average 62 years old but 82 years was counted for the two oldest individuals. It may even be possible to find remnant stems on the ground to extend our chronology to earlier periods through cross-dating with living individuals. The mean annual growth rate of snowy rhododendron is 0·36 mm and hence much wider than of dwarf shrubs in subarctic Manitoba, Canada (Au and Tardif, 2007) or in the Norwegian mountains (Bär et al., 2007). Favourable years for radial growth were determined to be 1937, 1972, 1975, 1978, 1998, 2002 and 2007, and poor growth conditions were observed in 1948, 1962–1964 and 1980–1995.
The mean sensitivity of the standard ring-width chronology (0·12) is lower than that of a deciduous dwarf shrub, Hippophae rhamnoides, in the arid Qilian Mts (Xiao et al., 2007), but is similar to that of timberline Georgei fir (Abies georgei var. smithii) in the Sygera Mts in south-east Tibet, China (Liang et al., 2009). Over the common period 1959–2007, the standard ring-width chronology has an RBAR of 0·33 and an EPS value of 0·89 (17 stems included), indicating a common macro-environmental influence on growth.
Interestingly, the snowy rhododendron chronology is significantly correlated (r = 0·39, 1932–2006, P < 0·01) with a well-replicated and summer temperature-sensitive timberline Georgei fir ring-width chronology in the Sygera Mts (Liang et al., 2009), 200 km away from the snowy rhododendron site (Fig. 4). This gives a high degree of confidence in the reliability of the cross-dating of the snowy rhododendron ring-width series. The existence of a common growth pattern of timberline trees and shrubs throughout the study area was also reported for the Andes in north-central Chile (Barichivich et al., 2009) and for the Norwegian mountains (Bär et al., 2007). Similar to timberline Georgei fir (Liang et al., 2009), alpine snowy rhododendron growth also captures the recent warming in south-east Tibet, suggesting that its growth is a sensitive ecological indicator for global change.
The radial growth of snowy rhododendron is highly dependent on an above-average warm and moist preceding November (both the mean minimum temperature and the sum of precipitation are strongly positively correlated with growth-ring width); moreover, mean minimum temperature in the current July is positively correlated with radial growth (Fig. 5). Based on seasonal cambial activity data, Rossi et al. (2008) observed that the radial growth of timberline trees is related to minimum rather than to maximum temperature. Over a distance of around 800 km, from the arid northern Tibetan Plateau (Liang et al., 2006; Zheng et al., 2008; Zhu et al., 2008) through the semi-humid eastern Tibetan Plateau (Shao and Fan, 1999; Bräuning, 2006; Liang et al., 2008) up to the humid south-eastern Tibetan Plateau (Liang et al., 2009), it has been reported that the monthly mean minimum temperatures in November of the previous year and in July of the current year are the dominant and unifying growth-limiting factors for timberline coniferous tree species. The mean minimum temperature in November may be one of the factors controlling the altitudinal distribution of snowy rhododendron. Vegetative buds, which are active in autumn, may become damaged during cold nights in early winter (Taschler and Neuner, 2004). A difference of occasionally about 14 °C between days and nights in November may cause frost damage to leaves and buds and hence reduce root activity or increase the risk of frost-induced desiccation. Subsequent defoliation and bud mortality would deplete the reservoir of carbon and growth hormones and thus reduce the tree's potential for further growth and photosynthetic activity.
As expected, moisture availability is not problematic for alpine snowy rhododendron in the study area. Apart from precipitation in the previous November, no significant relationships between tree growth and precipitation were discerned. Moreover, no significant effect of November precipitation on the growth of timberline trees was found in south-east Tibet (Liang et al., 2009). Precipitation in November accounts for only 2·1 % of the annual total precipitation and there is no correlation between precipitation and the mean minimum temperature in November. November precipitation (in the form of snow) may protect the buds from frost; moreover, in a cloudy November, there may be fewer cold nights as outgoing long-wave radiation is hampered. Nevertheless, it deserves further study to explain the positive effect of precipitation on snowy rhododendron growth.
This study suggests a high dendrochronological potential for snowy rhododendron, a shrub species which was explored for the first time. Its growth rings are well-defined and cross-dating is possible; moreover, there is a clear climatic signal in the growth-ring variability. In particular, it displays a climate/growth association similar to timberline trees on the Tibetan Plateau, with a monthly mean minimum temperature of the previous November and of the current July being the unifying growth-limiting factors. This finding further strengthens the fact that the annual growth rings of alpine snowy rhododendron contain a far-reaching temperature signal. Its growth may benefit from global warming. Given a wide geographic distribution of rhododendrons on the Tibetan Plateau, it shows a highly promising potential for future dendrochronological work in those areas where trees cannot survive. However, longer ring-width series, possibly of 200 years, and a good site replication will be needed to improve our understanding of the climate/growth association of the alpine rhododendron. Climate reconstruction from alpine shrub growth rings may be used in conjunction with other proxy data (e.g. ice cores and lake sediments) (Yao et al., 2002; Zhu et al., 2003; Tian et al., 2006) and ecological indicators (Luo et al., 2004) in future studies on the Tibetan Plateau. In addition, the dendrochronological potential of rhododendron is inspiring the exploration of the variation in radial growth of more rhododendron species occurring in various forms from subtree to shrub and dwarf shrub along altitudinal gradients of their distribution and their climatic responses in south-east Tibet.
We thank Peter Brown and an anonymous reviewer for helpful comments on an earlier version of our paper. We also thank Yan Xu for preparing Fig. Fig.11 and Haiqing Xu for help in the field. This work was supported by the National Natural Science Foundation of China (grant numbers 30670383, 40871097); and the K. C. Wang Education Foundation, Hong Kong (to E.L.).