Environ Geol (2007) 53:543–551 DOI 10.1007/s00254-007-0666-1
O R I G I N A L A RT I C L E
Radial growth of Tamarix ramosissima responds to changes in the water regime in an extremely arid region of northwestern China Shengchun Xiao Æ Honglang Xiao
Received: 2 December 2006 / Accepted: 29 January 2007 / Published online: 13 February 2007 Springer-Verlag 2007
Abstract The response of radial growth of tamarisk (Tamarix ramosissima) growing on the shore of WestJuyan Lake, on the Heihe River in northwestern China, to changes in the lake’s water regime was studied using tree-ring chronologies, principal components (PC) analysis, and classical correlation analysis. The first PC accounted for 53.3% of the total variance and reflected a common growth response at different sites. Correlation analysis indicated that fluctuations in the lake’s water level during the growing season (May–August) was primarily responsible for variations in the radial growth of tamarisk and explained more of the variance at low-lying sites than at higher sites. The second PC accounted for 30.7% of the total variance and revealed distinct differences in growth response between low-lying sites and those on higher ground. Total annual precipitation played an important role in radial growth of tamarisk at the higher sites. The spatial pattern in the tree-ring chronologies for different sites was performed in the temporal pattern of the treering chronology at the same site. Other factors such as microtopography, soil salinity, sand activity, and browsing by herbivores also affected the radial growth of tamarisk. The diversity in responses to the maximum water table depth for tamarisk in the study area appears to have been caused by local variations in precipitation, which can compensate to some degree for the inability of a plant’s roots to reach the water table.
S. Xiao (&) H. Xiao Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Donggang West Road, 320, 730000 Lanzhou, China e-mail:
[email protected]
Keywords Ecological response Dendrohydrology Water regime changes Tamarisk Water table fluctuation
Introduction Variation in a site’s water regime is a main factor that influences the composition and pattern of the natural vegetation in oases in arid regions. Changes in water regime have been implicated in determining the pattern of the natural vegetation that develops and in the vegetation’s growing conditions (Zhao and Cheng 2001). In arid regions, the depth of the water table deeply affects the soil moisture content in different layers of the soil, especially in the main layers where the root system is distributed as well as the growth, development, distribution, and succession of plants at a site (Zhao et al. 2003a). The ecological water table depth (EWTD) includes the range of water table depths that can maintain water consumption and normal growth by the natural vegetation (Zhao et al. 2003b). The lowest water table depth that satisfies this criterion is referred to as the maximum EWTD. At greater depths, water cannot move into the upper soil layers by means of capillary action and thus cannot become available for plant roots to absorb, and the resulting dryness of the soil leads to deterioration of the vegetation community (Feng et al. 1998). A study in the lower reaches of the Shiyang watershed in the Hexi Corridor of northwestern China showed that the percent coverage of sites by tamarisk (Tamarix ramosissima) as a function water table depth resembled a parabola and that the growth
123
544
of tamarisk was affected mainly by variations in the water table (Yang and Gao 2000): the growth of the tamarisk increased as the water table declined to a depth of 9–10 m, then declined gradually as the depth to the water table increased beyond this depth. Song et al. (2000) investigated the relationship between the depth of the water table and the frequency of tamarisk appearance in the Talimu watershed and found that about 91% of the tamarisk was found where the depth to water was less than 7 m. In the lower reaches of the Heihe watershed, the maximum EWTD of tamarisk is within 7–8 m of the surface (Feng et al. 1998). The dominant species of natural vegetation found in oases in arid regions rely most strongly on soil water to support evapotranspiration and depend relatively little on precipitation (Song et al. 2000). For phreatophytes such as tamarisk, variation in the water table is the main limiting factor to their growth and development. However, the observed variation in the coverage of a site by tamarisk as a function of depth to water suggests that other environmental factors might also affect the growth of tamarisk. Another question is raised by the very different EWTD values for tamarisk in the lower reaches of three different watersheds: despite strikingly similar water regimes, tamarisk growth varies among the study sites. Plant growth and its zonation patterns result from hydrological processes on large temporal scales. In arid regions, these processes include the effects of surface and soil water, but also the effects of limited precipitation. A study of isotopic variation in the xylem cellulose of tamarisk showed that the water source for saplings was mainly local precipitation, whereas young and mature trees mainly utilized soil water to support root system development (Yang et al. 1996). Although precipitation is scarce in arid regions, it has the un-substitutable functions on reducing the evapotranspiration of vegetation, shallow water, and soil. Wu et al. (2002) estimated supplemental inputs of soil water by precipitation amounting to 1.7 · 108 m3 per year in the Ejina Basin. However, there have been no studies of both the response of tamarisk to changes in the water regime and the contribution of the local precipitation, particularly in areas where stream flow decreased gradually over time and the soil water became a limiting factor. In the present study, we tried to answer these questions using ring-width chronologies for tamarisk growing at different distances above the shore of China’s West-Juyan Lake.
123
Environ Geol (2007) 53:543–551
Study area, materials, and methods Site description The study area is located in Ejina Banner, Inner Mongolia, one of China’s extremely arid regions. Based on local meteorological station, the mean annual temperature (1957–2002) is about 8.2C (with mean temperatures of 26.3C in July and –12.2C in January), and total precipitation averages less than 50 mm, of which 84% occurs during the growing season (May– September). Potential evapotranspiration is greater than 3,500 mm. The dominant woody species are Populus euphratica and T. ramosissima (saltcedar). P. euphratica is the dominant native woody species along the Ejina River. However, saltcedar is a pioneer species that can form monospecific stands on high river terraces and on lake shores. The Heihe River, with an 821-km-long main stream, originates in the middle of the Qilian Mountains and flows through the Hexi Corridor of Gansu Province, ultimately draining a watershed of 1,30,000 km2 before terminating at the northern end of East-Juyan and West-Juyan lakes in Ejina Banner, Inner Mongolia. Increasing demand for the water resources provided by oases in the Hexi Corridor in recent decades has transformed the lower reaches of the Heihe River (also referred to locally as the Ejina River) from a perennial river into an ephemeral stream. Our study was carried out at the southwestern margin of Lake West-Juyan (4220¢15.56†–4221¢8.81†N and 10041¢50.945†–10042¢30.63†E) (Fig. 1). Three clear beach terraces on the shore extend for several kilometers: sequentially, these are named B1, B2, and B3 moving from nearest to farthest from the lake shore. The slope gradient increases from 2.5 m per 1,000 m of horizontal distance near the lake to 32.1 m at the B3 terrace. The beach soil is a loose saline-alkaline takyr, which is always covered with a salt crust. The salt crust is thicker than 20 cm on the lowest portion of the lake floor center. The vegetation community is dominated by T. ramosissima, but the dwarf shrub Reaumuria soongorica is common at higher elevations above the shore. The vegetation changes from a dwarf Reaumuria community (above the B3 beach terrace) and a Reaumuria–Tamarix community (between B3 and B2, a distance of about 90 m) to a pure Tamarix community (between B2 and B1). The saltcedar cones range in height from 2 to 3.5 m between B3 and B2 and below 2 m at B2. A thin layer of shrub litter exists under the shrubs at the B1 beach terrace and the saltcedar had not
Environ Geol (2007) 53:543–551
545
Fig. 1 Locations of the study area and the site on the shore of West-Juyan Lake, one of the terminal lakes of the Heihe River of Inner Mongolia, China
formed sandy cones. The saltcedar grow very weakly toward the middle and top of the north-facing slope of the B2 beach terrace, with an average height of 2 m above the top of the sandy cones. The saltcedar shrubs grow strongly on the north-facing slope of the B1 beach terrace, with an average height of 2–2.5 m (Xiao et al. 2004). Sampling strategy For each shrub, two to three of what appeared to be the oldest branches were selected for sampling. Wood discs were collected at the base of each sampled Table 1 Characteristics of the Tamarix ramosissima ring-width chronologies at three beach terraces at West-Juyan Lake in Inner Mongolia, China
a
Ranges in parentheses
branch. Where saltcedar cones had formed, the discs were cut at the surface of the cones. From the lake floor to the highest terrace, 73 samples were collected, of which 21 were collected on the north-facing slope of B1, 28 (including 5 dead samples) were collected from B2, and 24 (including 3 dead samples) were collected from B3. The elevations of the three sampling sites were determined by means of GPS positioning (Magellan System Corp., San Dimas, CA, USA) with a precision of up to (1 cm per meter of elevation; the resulting elevations ranged from 893.2 to 904.3 m a.s.l., and the corresponding relative heights above the lake floor ranged from 3.8 to 16.1 m (Table 1).
Terrace
Elevation (m a.s.l.) Height of sampling site above lake floor (m) Chronology length (years)a Trees (N) Mean sensitivity Serial correlation Signal to noise ratio First-order autocorrelation Cumulative contribution of the first two principal components (%)
B1
B2
B2–B3
893.2 3.8
896.8 8.6
904.3 16.1
45 (1958–2002) 21 0.265 0.57 6.37 0.50 61.8
100 (1903–2002) 28 0.217 0.36 5.45 0.66 35.1
176 (1827–2002) 24 0.18 0.34 2.03 0.67 27.0
123
546
Development of the ring-width chronology In the laboratory, the discs were polished with sandpaper of progressively finer grit. Ring widths were measured with a Lintab tree-ring measurement system to the nearest 0.01 mm in two directions. A few samples were measured in only one direction because of their strong eccentricity and narrow growth zone. The measured tree-ring sequences were cross-dated and the quality of the analysis was assessed using the COFFCHA software (Holmes 1983). The ring-width chronology for the three sampling sites was developed using the ARSTAN software (Grissino-Mayer et al. 1996). The length of the chronology used for our analysis was chosen such that the interval had at least six sample replications. Chronological statistics such as the mean sensitivity (a measure of the annual variability in tree rings), the serial correlation (a measure of the amount of common signals among tree-ring sequences), the signal to noise ratio (a measure of the strength of the common signal relative to uncommon signals that represent noise), and the auto-correlation (a measure of the association between growth in the previous year and the current year) were obtained to show the characteristics of the tree-ring chronologies. Principal components analysis of the site chronologies The tree-ring chronologies at the three beach terraces exhibited responses to a variety of local and regional environmental signals. Principal components analysis (PCA; Zhang and Hebda 2004; Brubaker 1980) was used to summarize the regional variation in radial growth patterns revealed in the chronologies for the three terraces. This analysis was conducted using the PCA software (Grissino-Mayer et al. 1996) and the correlation matrix for the three site chronologies for the common interval from 1958 to 2002. The weight associated with each chronology conveyed information about the characteristic relationship between growth at each site and the PC: the higher the weight, the closer the relationship. Relationships between hydrology, climate, and growth The water status of the soil is the main factor that limits plant growth in extremely arid regions. Therefore, our analysis correlated growth of saltcedar with local hydrological conditions and the regional climate record obtained from the nearest observation stations.
123
Environ Geol (2007) 53:543–551
The water table data was obtained from the Ejina Banner Department of Water Resources using records from 1995 to 2002. The well used to collect this data was 15 km from the study site at Saihannur in the lower reaches of the West Ejina River. Climate data was obtained from the nearest Ejina Banner meteorological station, 60 km from the sampling site, and consisted of monthly precipitation and total annual precipitation from 1957 to 2002. The locations of these sites are shown in Fig. 1. A previous study suggested that radial growth of the tamarisk at the lake shore was mainly controlled by variations in soil water during the growing season (Xiao et al. 2004). The present study included both the local environmental conditions at the tamarisk stands and the regional climatic data. Regression analysis was used as an empirical method of extracting the hydrologic and climatic signals from each chronology. A preliminary analysis of tamarisk growth, local hydrology, and regional climate was performed for the three chronologies, each of which represented tamarisk growth along a gradient in relative height above the lake.
Results Tree-ring chronologies Ring-width chronologies for tamarisk were developed for each of the three beach terrace sites (Fig. 2). The results demonstrate that the chronologies exhibited similar variations during the period of overlap despite the fact that they came from trees growing at different terraces. The predominantly strong agreement between chronologies indicates that the growth of the trees at all three terraces was influenced by the same environmental conditions. The chronological statistics, which present variety in the spatial pattern among the three beach terraces, are presented in Table 1. The mean sensitivity ranged from 0.18 to 0.26, indicating that the ring-width variability was closely related to the gradient in relative height: the higher the terrace, the lower the mean sensitivity. The mean serial correlation (0.34–0.57), the signal to noise ratio (2.03–6.37), and the cumulative contribution of the first two principal components (27.0–61.8%) followed the same trend, suggesting that the individual trees at each site contained a sufficiently common environmental signal in their growth rings. The values of the first-order autocorrelation (0.50–0.67) followed the opposite trend, indicating that the chronologies contained low-frequency variance generated by a lag effect resulting
Environ Geol (2007) 53:543–551
547
Fig. 2 Ring-width chronologies for Tamarix ramosissima at three beach terraces above West-Juyan Lake in Inner Mongolia, China. Black strips represent periods of strong agreement in the variation during intervals common to the three chronologies
from the local water table, regional climate, and tree physiology in the current year on growth in the following year. In the temporal pattern of the same chronology, fluctuations in the B2–B3 site chronology decreased after the 1890s, but decreased after the early 1960s in the B2 site chronology. This phenomenon, however, did not appear in the B1 chronology, which showed relatively constant values from about 1870 onward. Radial growth patterns as a function of terrace elevation The PCA of the three site chronologies showed that the first two PCs accounted for 53.3 and 30.7% of the total variance, respectively, for the period of overlap (1958– 2002; Table 2). The third PC accounted for less than 16% of the total variance that described other variability, and was thus considered insignificant. The first
Table 2 Principal components of the Tamarix ramosissima chronologies (1958–2002) at three beach terraces above WestJuyan Lake in Inner Mongolia, China Principal component
Eigenvalue
Variance (%)
Cumulative variance (%)
PC1 PC2 PC3
1.60 0.92 0.48
53.3 30.7 16.0
53.3 84.0 100
two PCs were thus used in the subsequent analysis to examine the spatial patterns in the tamarisk tree rings. The weights associated with the chronologies were plotted against the three terraces to display the spatial patterns of growth represented by the first two PCs (Fig. 3). For PC1, the weights of the three chronologies all had positive values. This pattern indicates that the growth variations at the three terraces were positively correlated with PC1. Therefore, the first PC reflected a common growth response for all three terraces, with a larger contribution from the lower terraces. For PC2, the weight for the B1 terrace was negative, and the weights for the B2 and B2–B3 terraces were positive. This observation indicates that the component of radial growth represented by PC2 was positively correlated with growth at the higher terraces (the B2 and B2–B3 sites) and negatively correlated with growth at the lower terrace (the B1 site). Overall, PC2 reflected a pattern of decreasing impact of water table depth with increasing distance from the lake. In order to identifying the contributions of the PCs to the chronologies, the relationships between the PCs and the ring-width chronologies on the three beach bars were analyzed (Table 3). It showed that the PC1 have a positive significant correlation to the chronologies in three beach bars, and PC1 present the common changes of them. Thus, the weight values of the chronologies are all positive to the PC1, and so are the PC2 and PC3. The growth variations represented by the first two PCs were displayed by plotting the PC scores against the corresponding calendar years (Fig. 4). The most striking features of PC1 were that the largest and most positive scores were observed from 1996 to 1999 and from 1979 to 1981, respectively; the largest and most negative scores were observed from 1958 to 1960 and from 1983 to 1986, respectively. The large positive scores indicate above-average growth throughout the study area in those years. Conversely, the large negative scores indicate below-average growth throughout the study area. This spatial synchronicity in the growth pattern can also be seen by directly comparing the observed growth patterns at the three terraces (Fig. 2). Such a comparison reveals that common intervals of enhanced growth occurred from 1979 to 1982 and from 1996 to 1999; similarly, common intervals of reduced growth occurred from 1958 to 1962 and from 1983 to 1990 (Fig. 2). The scores for PC2 reveal variation in one or more environmental factors not explained by PC1 (Fig. 4). Positive values indicate above-average growth for sites with positive weight and below-average growth for sites with negative weight. Conversely, negative values
123
548
Environ Geol (2007) 53:543–551
uted strongly to tamarisk survival when the depth to the water table increased so much that the root systems of the tamarisk could not reach this water.
Weight value
1 PC1 0 PC2
Discussion and conclusion -1
B1
B2 sampling sites
B2-3
Fig. 3 Weights associated with the first (PC1) and second (PC2) principal components of the tree-ring chronologies for Tamarix ramosissima at three terraces above West-Juyan Lake in Inner Mongolia, China
mean that the sites with negative weights had aboveaverage growth and sites with positive weights had below-average growth. Relationships between water regime and growth To clarify the variation in the response of saltcedar to water regime (including soil water and precipitation) at the three terraces, we calculated the correlation of the first two PCs with the depth to the water table and with total precipitation. Table 4 shows that PC1 was positively and significantly correlated (P < 0.05) with the mean depth to the water table during the growing season (May–August), and weakly and nonsignificantly positively correlated with total precipitation. In contrast, PC2 was positively and significantly correlated (P < 0.05) with total precipitation, and weakly and negatively correlated with the depth to the water table during the growing season. Neither PC was significantly correlated with any of the temperature data. This result suggests that variation in the depth of the water table was the dominant factor that determined radial growth of tamarisk when there was sufficient water for this species, but that precipitation contribTable 3 Correlation matrix among the PCs and the chronologies in the West-Juyan Lake
PC1
PC2
PC3
B2–B3
B2 **Correlation is significant at the 0.01 level (two-tailed)
123
Pearson correlation Sig. (two-tailed) N Pearson correlation Sig. (two-tailed) N Pearson correlation Sig. (two-tailed) N Pearson correlation Sig. (two-tailed) N Pearson correlation Sig. (two-tailed) N
Tree-ring growth characteristics over large-scale geographical regions often reveal substantial differences in growth between sites (Zhang and Hebda 2004; Gou et al. 2004; Yoo and Wright 2000), and constructing an exact tree-ring chronology for a site can provide insights into regional climatic or environmental variation. Our study showed that the radial growth of tamarisk at beach terraces at different elevations above a lake shore contained common signals related to both the depth of the water table and to regional precipitation: the lower the terrace, the stronger the effect of the water table on radial growth; the higher the terrace, the greater the contribution of local precipitation to radial growth. The water table of the lowest beach terrace (B1) was less than 5 m below the surface after the 1960s. The weight of the B1 site chronology was positive for PC1 and negative for PC2; that is, the B1 site chronology represented a strong contribution by PC1 but a weak contribution by PC2. The weights of the chronology for the B2 site, where the depth to the water table was 8.6–11.5 m after the 1960s, were positive for both PC1 and PC2, but the contribution of PC1 was larger. The weights for the B2–B3 site, where the depth to the water table was 16.1–18.9 m after the 1960s, were negative for PC1 but positive for PC2. These results suggest that fluctuations in the depth of the water table significantly affected tamarisk growing at the lower terraces, but weakly affected the tamarisk growing at higher terraces; in contrast, precipitation had a strong effect on tamarisk growing at the higher
PC2
PC3
B2–B3
B2
B1
0.000 1.000 45
0.000 1.000 45 0.000 1.000 45
0.458** 0.002 45 0.573** 0.000 45 –0.108 0.481 45
0.860** 0.000 45 –0.068 0.656 45 –0.506** 0.000 45 0.409** 0.005 45
0.755** 0.000 45 –0.502** 0.000 45 0.422** 0.004 45 0.013 0.933 45 0.470** 0.001 45
Environ Geol (2007) 53:543–551
549
Fig. 4 The first and second principal components for the Tamarix ramosissima ring-width chronologies at three terraces above West-Juyan Lake in Inner Mongolia, China
Table 4 Result of the correlation analysis for the first two principal components (PCs) and local records for the water table depth (1995–2002) and regional data for total precipitation (1957–2002)
PC1 PC2
Mean depth to the water table from May–August
Total precipitation
Correlation coefficient
Significance (two-tailed test)
Correlation coefficient
Significance (two-tailed test)
0.803* –0.125
0.016 0.769
0.068 0.310*
0.659 0.038
*Correlation is significant at P < 0.05
terraces, and little effect on tamarisk growing nearer to the lake. This conclusion reflects the ecological and spatial response patterns for the radial growth of the tamarisk. This ecological response pattern was also evident on a temporal scale. The depth to the water table (or the lake level before the lake began drying in 1963) was less than 7.5 m before the 1900s, ranged from 7.5 to 11.4 m from 1900s to the 1950s, and ranged from 16.1 to 18.9 m after the 1960s. Therefore, the amplitude of the fluctuation in the B2–B3 site chronology was greatest before the 1890s and smallest thereafter (Fig. 2). At the B2 site, the depth to the water table was less than 4.8 m from the 1900s to the 1950s, and ranged from 8.6 to 11.5 m after the 1960s. As a result, the B2 site chronology revealed two different periods:
before and after the 1960s (Fig. 2). However, there was no decrease in the amplitude of the fluctuations in the B1 site chronology because the depth to the water table remained less than 5.0 m since the tamarisk became established (Fig. 2). The changes in the amplitudes of the fluctuations in the three site chronologies reveals that the radial growth of tamarisk at the terrace sites was mainly controlled by the depth to the water table when the maximum EWTD was no more than 5 m. However, other factors, such as total precipitation, could also greatly affect the radial growth of tamarisk. This result reflects the temporal ecological response pattern of the radial growth of tamarisk to the first two PCs as the lake level declined and precipitation became a more important component of a plant’s total available water resource. A previous study in the same region (Zhou et al. 2004) indicated that the evapotranspiration rates of tamarisk would decrease significantly on cloudy and rainy days. In addition, Anderson (1982) found that relative humidity directly affected the evaporation effect and, thus, the water potential of tamarisk tissues. These results suggest that solar radiation is the limiting factor for tamarisk growth when there is sufficient water. Increased numbers of cloudy and rainy days lead to lower overall solar radiation, which reduces photosynthetic rate and thus reduces the radial growth of tamarisk. This explains why the B1 site chronology had a negative weight for PC2. In contrast, increasing solar insolation would increase the water stress of tamarisk when water is the primary limiting factor. High temperatures and strong radiation input intensify evaporation rates, further decreasing moisture content in the top layers of the soil. Under these conditions, an increased frequency of cloudy and rainy days would mitigate water stress and promote the radial growth of tamarisk. This is why the B2 and B2–B3 site chronologies had positive weights for PC2. Anderson (1977) found that tamarisk is a phreatophyte that can reach the water table with its deep roots. After analyzing the changes of soil water contents and the distribution of roots in tamarisk stands before and after irrigation, Zeng et al. (2002) deduced that the increased transpiration of tamarisk after irrigation resulted from increased utilization of deep soil water, but did not result from increased water content in the upper layer of the soil. These results suggest that variation in the depth to the water table was the primary limiting factor, which supports the results of our own study, in which variation in the depth of the water table was also significantly correlated with PC1. PC2 was significantly and positively correlated with total precipitation. This suggests that the contribution
123
550
of total precipitation to a plant’s water balance would increase and could compensate for shortages in soil water to some extent when the depth of the water table increased beyond the depth at which tamarisk can exploit this water. Good growth of tamarisk could persist longer due to the contribution of local precipitation, rather than declining sharply as the water table receded. The growth of tamarisk would then decrease only when precipitation could not compensate for the water shortage induced by the receding water table. This provides a good explanation for the parabolic shape of the curve for the relationship between tamarisk growth and the depth to the water table reported for tamarisk in the lower reaches of Shiyang watershed (Yang and Gao 2000). The maximum EWTD is within 10 m of the surface in that area. Studies conducted in the lower reaches of the Heihe and Talimu watersheds (Feng et al. 1998; Song et al. 2000) indicated that tamarisk grows vigorously when the depth of the water table is less than 5 m, whereas the trees survive but do not thrive when this depth is 5–7 m, and exhibit growth reductions or mortality of most individuals at water table depths greater than 7–8 m. The mean total precipitation was less than 42, 38.2, and 108.9 mm in the lower reaches of the Talimu, Heihe, and Shiyang watersheds, respectively, based on local climatic data. Therefore, the maximum EWTD could extend up to 10 m in the lower reaches of the Shiyang watershed, where precipitation compensated for a lack of water, but could only extend to 7–8 m in the Talimu and Heihe watersheds, where precipitation was much lower. The differences among regions in EWTD that were reported in the literature thus reflect the degree to which local precipitation can compensate for water shortages induced by the receding water table. A decline in the tamarisk community, composed of phreatophytes, is a signal that the local water table has receded too far for the local precipitation to compensate, and that the local water environment has deteriorated strikingly to the point at which it may be unable to maintain shrub growth and a stable vegetation community. The cumulative contribution of the first two PCs was 84%. Therefore, factors other than the depth to the water table and precipitation have some effect on tamarisk growth; these may include microtopography, soil salinity, and disturbance by wind-blown sand and grazing. He et al. (2003) reported that the formation, development, and degeneration of a Tamarix taklamakanensis community may have been affected by windblown sand. In our study area, soils, salinity, and the intensity of livestock grazing determine the formation of
123
Environ Geol (2007) 53:543–551
tamarisk sand cones. The ecological response patterns of tamarisk to these factors require further research. In conclusion, the depth of the water table appears to be the major factor of the water environment that limits the growth of tamarisk when this depth is within the range of values for EWTD suitable for tamarisk growth in the arid regions of northwestern China. Precipitation contributed greatly to tamarisk growth when the water table receded sufficiently far below this range of depths that the receding water table adversely affected growth. As a result, the maximum EWTD varied among arid regions based on the contribution of local precipitation to plant water availability. Acknowledgments This research was National Natural Science Foundation No. 40371009) and the projects which Science and Technology Ministry (Grant 2006BAD26B02).
supported by the of China (Grant initiated by China No. 2005BA517A03;
References Anderson JE (1977) Transpiration and photosynthesis in saltcedar. Hydrol Water Resour Arizona Southwest 7:125–131 Anderson JE (1982) Factors controlling transpiration and photosynthesis in Tamarix Chinese Lour. Ecology 63:46–50 Brubaker LB (1980) Spatial patterns of tree growth anomalies in the Pacific Northwest. Ecology 61:798–807 Feng Q, Cheng GD, Tan ZG (1998) Study on the relationship between the vegetation growth and ecological groundwater table in oasis. J Desert Res 18(Suppl 1):106–109 Gou XH, Chen FH, Yang MX, et al (2004) Analyses of the treering width chronology of Qilian Mountains at different elevation. Acta Ecol Sin 24:172–176 Grissino-Mayer HD, Holmes RL, Fritts HC (eds) (1996) The international tree-ring data bank program library, user’s manual, version 2. International Tree-Ring Data Bank, Tucson He XD, Gao YB, Ren AZ (2003) Role of wind-sand disturbance in the formation and development of Tamarix taklamakanensis community. Acta Bot Sin 45:1285–1290 Holmes RL (1983) Computer-assisted quality control in tree-ring dating and measurement. Tree Ring Bull 43:69–78 Song YD, Fan ZL, Lei SD, et al (2000) Study on the problems of water resources and ecology in Talimu watershed, China. Xinjiang People’s Press, Wulumuqi Wu XM, Shi SS, Li ZH, et al (2002) Study on the groundwater flow system in Ejina basin of the lower reaches of Heihe river, northwest China (first part). J Hydrogeol Eng Geol 1:16–20 Xiao SC, Xiao HL, Zhou MX, et al (2004) Water table change of the West Juyan Lake in the past 100 years recorded in the tree ring of the shrubs in the lake banks. J Glaciol Geocryol 26:557–562 Yang ZH, Gao ZH (2000) Impact of precipitation and undergroundwater table in the edge of oases on growth and decline of Nitraria tangutorum community. Chin J Appl Ecol 11(6):923–926 Yang WB, Spencer RJ, Krouse HR (1996) Stable sulfur isotope hydrogeochemical studies using desert shrubs and tree rings,
Environ Geol (2007) 53:543–551 Death Valley, California, USA. Geochim Cosmochim Acta 60:3015–3022 Yoo SJ, Wright BD (2000) Persistence of growth variation in tree-ring chronologies. For Sci 46:507–520 Zeng FJ, Foetzki A, Li XY, et al (2002) A preliminary study on the effect of irrigation on eater physiology of Tamarix ramosissima in Cele Oasis. Chin J Appl Ecol 13:849–853 Zhang QB, Hebda RJ (2004) Variation in radial growth patterns of Pseudotsuga menziesii on the central coast of British Columbia, Canada. Can J For Res 34:1946–1954 Zhao WZ, Cheng GD (2001) Review of several problems on the study of eco-hydrological processes in arid zone. Chin Sci Bull 46:1851–1857
551 Zhao GD, Liu SR, Ma QL (2003a) Ecophysiological responses of two xerophytes Atraphaxis frutescens and Elaeagnus angustifolia to the change of groundwater depth in arid area (I). Acta Phytoecol Sin 27:228–234 Zhao WZ, Chang XL, He ZB (2003b) The response of the riparian forest distribution to the hydrological process in Ejina Oasis. Sci China Ser D 33(Suppl):21–30 Zhou MX, Xiao HL, Zhang XY, et al (2004) Preliminary analysis on evapotranspiration of Saltcedar community in Ejina Oasis. J Desert Res 24:479–483
123
