Relating Water Use to Tree Vitality of Populus euphratica Oliv . in the Lower Tarim River , NW China

This study aimed to compare the hydraulic characteristics of different vitalities of Populus euphratica to reveal the differences in their water use strategies and water consumption to provide useful data to scale water use of riparian poplar forests in the lower reaches of the Tarim River, Northwestern China. Our results showed that the sapwood area of P. euphratica could be estimated based on its correlation with tree biometric parameters. The sapwood area of vital poplars tended to be larger than the senesced poplar despite both having the same diameter at breast height. This indicates that poplar vitality should be taken into account when estimating its sapwood area. Therefore, we established two different sapwood area estimation models for vital and senesced poplar (sapwood area = 1.452 × DBH1.553, R2 = 0.891; sapwood area = 0.915 × DBH1.618, R2 = 0.718; DBH: diameter at breast height). The sap flow process of vital and senesced poplar had certain differences and similarities; the average diurnal sap flow velocity and water consumption of vital poplar were 15.85 cm/h and 45.95 L, respectively; for the senesced poplar, it was 9.64 cm/h and 18.17 L, respectively, which were smaller than that of vital poplars. The influence of environmental factors on the sap flow velocity of two different P. euphratica was similar; the sap flow of both vital and senesced poplar had positive correlation with air temperature (R2 = 0.800 and 0.851), solar radiation (R2 = 0.732 and 0.778), vapor pressure deficit (R2 = 0.508 and 0.643) and groundwater depth (R2 = 0.301 and 0.171), while negative correlation with air humidity (R2 = −0.313 and −0.478).


Introduction
Heterogeneous environmental conditions influence the hydraulic characteristics of vegetation.Water availability is one of the most important factors influencing the growth and development of plants [1].Related studies have indicated that the structure of the plant hydraulic system has the potential to limit water flow through its body, thus restricting its water balance, and eventually its growth [2].Therefore, studying the differences in the hydraulic characteristics of plants of different vitality may help us understand the influence of water availability on tree growth.
The Tarim River, located along the northern rim of the Taklamakan Desert in the Xinjiang Uighur Autonomous Region, Northwestern China.It is one of the longest inland water ways in the world along with the Volga, Syr Darya, Amu Darya, and the Ural [3][4][5][6].The riparian forests at the Tarim are an important ecosystem which plays a key role in maintaining the structure, function and stability of the arid ecosystem [7,8].
Populus euphratica Oliv. is a rare, ancient, and endangered tree species which forms the floodplain forest ecosystem in inland river basins in central Asian arid regions [9,10].Given its high capacity against environmental stressors, e.g., saline, high temperature, drought and sand storms, it plays an irreplaceable function in these areas [11][12][13][14].However, due to the irrational water allocation along the upper and middle reaches of the Tarim River, the lower reaches (ca.320 km) has been cut off since 1970.This has resulted in a decrease in the groundwater level along the lower Tarim [4,15,16], and the eventual degradation of the riparian ecosystem [6,12,17,18].Since 2000, the Ecological Water Diversion to the lower Tarim River was started to restore the degraded ecosystem of this region, and a number of studies have shown that the riparian forest has demonstrated a positive response to the water diversion [12,13,19,20].
However, in terms of water shortage in this region, the insurance of long-term water transferring and the virtuous cycle of ecological environment depends on the rational and efficient usage of limited water resources, which includes the scientific allocation of water for irrigated agriculture, nature and ecosystem conservation, as well as for industry, oasis settlements and human well-being [4,[21][22][23][24].Therefore, it is important to estimate the water use of riparian forests in the lower Tarim River for providing scientific reference to the water management authorities.Different upscaling methods have been reported for estimating the water use of poplar forests [25][26][27]; however, these methods did not consider the vitality of poplar trees.
The objectives of this study were to assess and compare the hydraulic characteristics of Populus euphratica of different vitalities by monitoring the diurnal changes of its sap flow during hot sunny days, measuring and estimating the area of water conducting tissue, as well as analyzing the correlation between sap flow velocity and environmental factors.These experiments were carried out on poplar trees of different vitalities growing at two sites with different groundwater conditions.There is evidence from the literature that the sapwood area of the poplar tree can be estimated based on the correlation between the sapwood area and tree biometric parameters [25,27]; furthermore, the water consumption of poplar trees at stand level could be obtained by upscaling methods based on the sap flow measurement of sample trees [26,28], but little information is available regarding the water consumption of poplar trees at different vitalities.This knowledge is of particular importance when estimating whole tree water usage in this region more accurately.

Sample Selection and Measurement
Based on the vitality classification criteria of Halik et al. [12] and Aishan et al. [20], Populus euphratica of different vitality status were chosen.Vital poplar sample trees were located at 40°08′44″ N, 88°20′59″ E; senesced poplar trees were located at 40°08′07″ N, 88°21′44″ E. Three sample trees were selected from each plot to run sap flow measurements.Sapwood samples were taken from 91 vital P. euphratica and 78 senesced P. euphratica using an increment borer, in addition, their biometric parameters (DBH: Diameter at breast height, CD: Crown diameter and TH: Tree height) were measured to run modeling analysis for estimating the sapwood area and model validation.Table 1 shows a summary of the statistics of the poplar trees used for sapwood sampling.Note: DBH = diameter at breast height, TH = tree height, CD = crown diameter.

Sample Selection and Measurement
Based on the vitality classification criteria of Halik et al. [12] and Aishan et al. [20], Populus euphratica of different vitality status were chosen.Vital poplar sample trees were located at 40 • 08 44" N, 88 • 20 59" E; senesced poplar trees were located at 40 • 08 07" N, 88 • 21 44" E. Three sample trees were selected from each plot to run sap flow measurements.Sapwood samples were taken from 91 vital P. euphratica and 78 senesced P. euphratica using an increment borer, in addition, their biometric parameters (DBH: Diameter at breast height, CD: Crown diameter and TH: Tree height) were measured to run modeling analysis for estimating the sapwood area and model validation.Table 1 shows a summary of the statistics of the poplar trees used for sapwood sampling.To calculate the sapwood area of the poplar stem, we assumed that the tree stem was circular, that the sapwood and the inner wood areas were in the shape of concentric circles, and the outer circle represented the sapwood area (Figure 2a).The sapwood of P. euphratica is easy to differentiate from the inner wood when the sample is fresh (Figure 2b), and can usually be observed with the naked eye.
To calculate the sapwood area of the poplar stem, we assumed that the tree stem was circular, that the sapwood and the inner wood areas were in the shape of concentric circles, and the outer circle represented the sapwood area (Figure 2a).The sapwood of P. euphratica is easy to differentiate from the inner wood when the sample is fresh (Figure 2b), and can usually be observed with the naked eye.The following equation is for calculating the sapwood area: where π is the mathematical constant (3.14), and d is the diameter of tree at breast height.

Sap Flow Measurement
A SFM1 sap flow meter (ICT, Armidale, NSW, Australia) that utilized the heat ratio method (HRM) was used to record heat pulse velocity (Vh) at 15 min intervals for the period from 1 to 10 September 2015.The installation and sensor design were identical to descriptions provided in Pfautsch et al. [31] and Keyimu et al. [32].

Meteorological Data
Meteorological parameters, such as Rs (solar radiation), Ta (air temperature), RH (air humidity) were collected from the corresponding sensors (Watchdog 2000, Spectrum Co. Ltd., Selmer, TN, USA).The parameters were recorded at 1 h intervals; data read out was completed using Spec ware software.Vapor pressure deficit (VPD) was calculated using the following equation [33]: where RH is air humidity, ea* is vapor pressure.ea* is obtained using the following equation [34]: The following equation is for calculating the sapwood area: where π is the mathematical constant (3.14), and d is the diameter of tree at breast height.

Sap Flow Measurement
A SFM1 sap flow meter (ICT, Armidale, NSW, Australia) that utilized the heat ratio method (HRM) was used to record heat pulse velocity (Vh) at 15 min intervals for the period from 1 to 10 September 2015.The installation and sensor design were identical to descriptions provided in Pfautsch et al. [31] and Keyimu et al. [32].

Meteorological Data
Meteorological parameters, such as R s (solar radiation), T a (air temperature), RH (air humidity) were collected from the corresponding sensors (Watchdog 2000, Spectrum Co. Ltd., Selmer, TN, USA).The parameters were recorded at 1 h intervals; data read out was completed using Spec ware software.Vapor pressure deficit (VPD) was calculated using the following equation [33]: where RH is air humidity, ea* is vapor pressure.ea* is obtained using the following equation [34]: where Ta is air temperature, and the values of parameters a, b, and c are 0.611 kPa, 17.502, and 237.3 • C, respectively.

Groundwater Data
The groundwater depth (GWD) measurement was carried out simultaneously with sap flow measurement using a Diver data logger (Schlumberger Limited, Paris, France) at two different sites.Logger installation, sensor design and measurement principles were provided in detail in Keilholz et al. [24,35].

Data Processing
A sap flow tool (ICT International, Armidale, NSW, Australia) was utilized to process the raw sap flow data, calculate the sap flow velocity, and water consumption.The number of trees and their biometric parameters data as well as sapwood area calculation were processed by Microsoft Excel 2015 software.Sapwood area estimation models for both vital and senesced P. euphratica were created by selecting the sapwood area as a dependent variable, and DBH, crown diameter, and tree height as the independent variables.Regression models between sap flow and meteorological parameters were established by selecting the VPD, RH, Ta, Rs, and GWD as the independent variables and sap flow rate as the dependent variable.Modeling analysis was carried out using the SPSS statistical tool 20.0 (IBM, Armonk, NY, USA).Paired t-test was conducted for validating the sapwood area estimation model at p ≤ 0.05 significance level.All plotting was carried out on Microsoft Excel 2015.

Sapwood Area of P. euphratica with Different Vitality
In the first plot, the results showed that DBH and sapwood area could be modeled at higher R 2 when compared to other biometric parameters (Figure 3), R 2 = 0.901, n = 48.In the meantime, the sapwood area of the rest of the trees was measured.The t-test result showed that there was no significant difference between the model's estimated value and the measured value (n = 43) of the sapwood area (p > 0.05) (Figure 4).Therefore, the estimation value of power function model was reliable, and could be used for estimating the sapwood area.
However, the performance result of the first model (which was established based on the biometric parameters of vital poplars) was not reliable when it was used to estimate the sapwood area of senesced poplars (Figure 5).As the estimated sapwood area tended to be larger than the measured sapwood area, the mean value of the model based estimated sapwood area and measured sapwood area were 166.04 cm 2 and 134.05 cm 2 , respectively.Therefore, another model was established using similar growth parameters of senesced poplars (Figure 6), R 2 = 0.718, (n = 40), the model validation result was reliable (p > 0.05) (Figure 7) (n = 38).
Water 2017, 9, 622 5 of 14 where Ta is air temperature, and the values of parameters a, b, and c are 0.611 kPa, 17.502, and 237.3 °C , respectively.

Groundwater Data
The groundwater depth (GWD) measurement was carried out simultaneously with sap flow measurement using a Diver data logger (Schlumberger Limited, Paris, France) at two different sites.Logger installation, sensor design and measurement principles were provided in detail in Keilholz et al. [24,35].

Data Processing
A sap flow tool (ICT International, Armidale, NSW, Australia) was utilized to process the raw sap flow data, calculate the sap flow velocity, and water consumption.The number of trees and their biometric parameters data as well as sapwood area calculation were processed by Microsoft Excel 2015 software.Sapwood area estimation models for both vital and senesced P. euphratica were created by selecting the sapwood area as a dependent variable, and DBH, crown diameter, and tree height as the independent variables.Regression models between sap flow and meteorological parameters were established by selecting the VPD, RH, Ta, Rs, and GWD as the independent variables and sap flow rate as the dependent variable.Modeling analysis was carried out using the SPSS statistical tool 20.0 (IBM, Armonk, NY, USA).Paired t-test was conducted for validating the sapwood area estimation model at p ≤ 0.05 significance level.All plotting was carried out on Microsoft Excel 2015.

Sapwood Area of P. euphratica with Different Vitality
In the first plot, the results showed that DBH and sapwood area could be modeled at higher R 2 when compared to other biometric parameters (Figure 3), R 2 = 0.901, n = 48.In the meantime, the sapwood area of the rest of the trees was measured.The t-test result showed that there was no significant difference between the model's estimated value and the measured value (n = 43) of the sapwood area (p > 0.05) (Figure 4).Therefore, the estimation value of power function model was reliable, and could be used for estimating the sapwood area.
However, the performance result of the first model (which was established based on the biometric parameters of vital poplars) was not reliable when it was used to estimate the sapwood area of senesced poplars (Figure 5).As the estimated sapwood area tended to be larger than the measured sapwood area, the mean value of the model based estimated sapwood area and measured sapwood area were 166.04 cm 2 and 134.05 cm 2 , respectively.Therefore, another model was established using similar growth parameters of senesced poplars (Figure 6), R 2 = 0.718, (n = 40), the model validation result was reliable (p > 0.05) (Figure 7) (n = 38).

Sap Flow Characteristics of P. euphratica with Different Vitality
There were certain similarities of sap flow variation between poplar trees of different vitality.Both vital and senesced poplars showed a pronounced diurnal sap flow process, all had peak variation features (Figure 8); and during the night the sap flow remained active, but at lower rate.
However, heterogeneous characteristics of sap flow variation also existed between the vital and senesced poplars.This was manifested from aspects of average sap flow velocity, daily consumed water amount, and increasing and decreasing time points of sap flow.The average sap flow velocity of vital poplars throughout the whole measurement period was 15.85 cm/h, which was higher than the average sap flow velocity of senesced poplar at 9.64 cm/h.The average daily consumed water amounts of vital and senesced poplars were 45.95 L and 18.17 L, respectively (Figure 9).Sap flow velocity of both vital and senesced poplars started to rise from 7:00 a.m., whereas the sap flow of senesced poplars reached its higher value of 13.69 cm/h at 10:00 a.m. and vital poplars reached its higher sap flow 25.68 cm/h at 12:00 p.m., after the midday depression appeared.From 17:00 p.m. onwards, the sap flow velocity of senesced poplars began to decrease, which was 2 h earlier than vital poplars.

Sap Flow Characteristics of P. euphratica with Different Vitality
There were certain similarities of sap flow variation between poplar trees of different vitality.Both vital and senesced poplars showed a pronounced diurnal sap flow process, all had peak variation features (Figure 8); and during the night the sap flow remained active, but at lower rate.
However, heterogeneous characteristics of sap flow variation also existed between the vital and senesced poplars.This was manifested from aspects of average sap flow velocity, daily consumed water amount, and increasing and decreasing time points of sap flow.The average sap flow velocity of vital poplars throughout the whole measurement period was 15.85 cm/h, which was higher than the average sap flow velocity of senesced poplar at 9.64 cm/h.The average daily consumed water amounts of vital and senesced poplars were 45.95 L and 18.17 L, respectively (Figure 9).Sap flow velocity of both vital and senesced poplars started to rise from 7:00 a.m., whereas the sap flow of senesced poplars reached its higher value of 13.69 cm/h at 10:00 a.m. and vital poplars reached its higher sap flow 25.68 cm/h at 12:00 p.m., after the midday depression appeared.From 17:00 p.m. onwards, the sap flow velocity of senesced poplars began to decrease, which was 2 h earlier than vital poplars.

Sap Flow Characteristics of P. euphratica with Different Vitality
There were certain similarities of sap flow variation between poplar trees of different vitality.Both vital and senesced poplars showed a pronounced diurnal sap flow process, all had peak variation features (Figure 8); and during the night the sap flow remained active, but at lower rate.
However, heterogeneous characteristics of sap flow variation also existed between the vital and senesced poplars.This was manifested from aspects of average sap flow velocity, daily consumed water amount, and increasing and decreasing time points of sap flow.The average sap flow velocity of vital poplars throughout the whole measurement period was 15.85 cm/h, which was higher than the average sap flow velocity of senesced poplar at 9.64 cm/h.The average daily consumed water amounts of vital and senesced poplars were 45.95 L and 18.17 L, respectively (Figure 9).Sap flow velocity of both vital and senesced poplars started to rise from 7:00 a.m., whereas the sap flow of senesced poplars reached its higher value of 13.69 cm/h at 10:00 a.m. and vital poplars reached its higher sap flow 25.68 cm/h at 12:00 p.m., after the midday depression appeared.From 17:00 p.m. onwards, the sap flow velocity of senesced poplars began to decrease, which was 2 h earlier than vital poplars.

Relationship between Sap Flow of Different Vitality P. euphratica and Meteorological Factor
The sap flow variation feature of vegetation was closely related to changing environmental factors during the measurement period [36].While conducting sap flow measurements on the sample trees, environmental factors were recorded simultaneously (Figure 10).The result of the non-parametric Spearman correlation analysis between sap flow velocity and environmental factors showed that the diurnal variation of sap flow velocity of both vital and senesced poplar was significantly related to VPD, Ta, RH, Rs and groundwater depth.It was negatively related to RH, and positively related to other meteorological factors (Table 2).The influence of meteorological factors on the sap flow rate of vital poplars was in the order of: Ta (0.800) > Rs (0.732) > VPD (0.508) > RH (−0.313) > GWD (0.301).For the senesced poplars, it was in the same order of: Ta (0.851) > Rs (0.778) > VPD (0.643) >RH (−0.478) > GWD (0.171).Note: ** Significant at 0.05 level.
The multiple regression equations showed that the main influencing factors on the sap flow rate of vital and senesced poplars were almost similar in the order of Ta, Rs, and GWD (Table 3).The regression coefficients of the equations were tested, and the significant error probability of the respective variables was <0.01, indicating that our regression equation revealed the interrelation between environmental factors and sap flow rate.

Relationship between Sap Flow of Different Vitality P. euphratica and Meteorological Factor
The sap flow variation feature of vegetation was closely related to changing environmental factors during the measurement period [36].While conducting sap flow measurements on the sample trees, environmental factors were recorded simultaneously (Figure 10).The result of the non-parametric Spearman correlation analysis between sap flow velocity and environmental factors showed that the diurnal variation of sap flow velocity of both vital and senesced poplar was significantly related to VPD, T a , RH, R s and groundwater depth.It was negatively related to RH, and positively related to other meteorological factors (Table 2).The influence of meteorological factors on the sap flow rate of vital poplars was in the order of: T a (0.800) > R s (0.732) > VPD (0.508) > RH (−0.313) > GWD (0.301).For the senesced poplars, it was in the same order of: T a (0.851) > R s (0.778) > VPD (0.643) > RH (−0.478) > GWD (0.171).The multiple regression equations showed that the main influencing factors on the sap flow rate of vital and senesced poplars were almost similar in the order of Ta, Rs, and GWD (Table 3).The regression coefficients of the equations were tested, and the significant error probability of the respective variables was <0.01, indicating that our regression equation revealed the interrelation between environmental factors and sap flow rate.

Discussion
A comparison analysis of the hydraulic characteristics of different plant species, the different individual of the same species (at different vitality), and the same species in heterogeneous environmental conditions should be based on their simultaneous observation and measurement; otherwise it does not make sense.
Sapwood is the water-conducting tissue [37][38][39], and direct correlation manifested between the sapwood area and water consumption of trees [27,40,41].Estimation of the sapwood area of a tree stand is commonly achieved based on the correlation between growth parameters of the tree (e.g., diameter at breast height, crown size, biomass, leaf area index) and the sapwood area.Hatton and Vertessy's study showed that the area of xylem conducting tissue, DBH, and tree basal area were efficient upscaling parameters [42,43].However, there are differences of sapwood width and sapwood area among different tree species, even among different individuals of the same tree species, which is influenced by the tree growth status [44,45], site condition [46,47], and tree genetic traits [48,49].Therefore, heterogeneous environmental conditions and the growth status of trees require different sapwood area estimation models, this support the findings in our study.

Discussion
A comparison analysis of the hydraulic characteristics of different plant species, the different individual of the same species (at different vitality), and the same species in heterogeneous environmental conditions should be based on their simultaneous observation and measurement; otherwise it does not make sense.
Sapwood is the water-conducting tissue [37][38][39], and direct correlation manifested between the sapwood area and water consumption of trees [27,40,41].Estimation of the sapwood area of a tree stand is commonly achieved based on the correlation between growth parameters of the tree (e.g., diameter at breast height, crown size, biomass, leaf area index) and the sapwood area.Hatton and Vertessy's study showed that the area of xylem conducting tissue, DBH, and tree basal area were efficient upscaling parameters [42,43].However, there are differences of sapwood width and sapwood area among different tree species, even among different individuals of the same tree species, which is influenced by the tree growth status [44,45], site condition [46,47], and tree genetic traits [48,49].Therefore, heterogeneous environmental conditions and the growth status of trees require different sapwood area estimation models, this support the findings in our study.
The diurnal sap flow variation feature of vital and senesced poplars in this study was similar as both showed obvious circadian rhythms; however, there were differences in the increasing and decreasing time points, as well as the average, maximum, and minimum value of sap flow.Average sap flow velocity of vital poplars was almost twice that of the senesced poplar.Sap flow of vital poplar started to decrease from 19:00 p.m., which was 2 h later than the senesced poplar, which may have been decided by the physiological and morphological status of the tree itself.Hörtensteiner's study showed that chlorophyll content in tree leaves decreased when the tree leaves were senesced [50]; furthermore, Christ et al. found that water shortage also induced decreasing chlorophyll in tree leaves [51], however, the lower chlorophyll content in tree leaves weakened the photosynthetic activity of tree leaves, which eventually reduced transpiration.Bucci et al. and Chave et al. presented studies that showed that lower water availability increased the density of sapwood and lowered the water transport capacity in trees [52,53].The average sap flow velocity during the day was higher than at night.Although the sap flow process was weaker at night, it remained active.This was probably to refill the water loss during the day to keep its water balance [54], which corresponded to the research results of Zhang et al. [55,56], Si et al. [57], and Yu et al. [58].
Water is the most influencing factor to the survival, distribution and development of vegetation in hyper arid regions [59].The root system is the main organ of the plant to absorb water from the ground, therefore, its morphology and distribution impacts the growth of vegetation.Feng et al. showed that P. euphratica had a highly developed root system, and its root density decreased with increasing distance from the ground [60], therefore, when the groundwater table was lower, the plant water uptake through its root would be restricted to some degree, and the amount of water transported by vegetation decreased.In contrast, a higher groundwater table enabled vegetation to uptake more water.Our findings on the difference of sap flow velocity and diurnal water consumption between two different measurement sites with heterogeneous groundwater depth corresponded with the research results of Ma et al. [61], Zhao et al. [62], and Shen et al. [63].
As demonstrated in our study, the groundwater depth was positively related to the sap flow velocity of both vital (R 2 = 0.301) and senesced (R 2 = 0.171) poplars.When the sap flow velocity increased, the groundwater depth also increased, which was due to the water uptake by poplar trees.The correlation of sap flow velocity and groundwater depth in the vital poplars distributed plot was larger than that of the senesced poplar plot, which indicates that the groundwater availability strengthened the influence on the water uptake of poplars.
Many studies have shown that Rs, RH, Ta, and GWD are the most influencing external factors on the water consumption of trees [40][41][42]56,64].Solar radiation affects the photosynthetic and transpiration activities of vegetation; it is the energy required to activate photosynthesis.Air temperature rises with the strengthening of solar radiation.When the leaf temperature is higher, plants decrease its temperature through increasing its transpiration rate.Leaf stomatal resistance increases when the air temperature reaches 32 • C, next, the stomata conductance and water potential decreases, which reduces the transpiration rate [65].Air humidity changes the VPD between the plant, leaf, and atmosphere.When the air humidity is low, the difference in vapor pressure between the leaf and atmosphere is higher, which accelerates the transpiration rate.In contrast, when the air humidity is higher, the transpiration rate is lower.

Conclusions
As reported in the relevant literature, the results of our study further showed that DBH and the sapwood area of poplars were closely related to each other; therefore, DBH could be used to estimate the sapwood area by establishing statistical models.However, the sapwood area of vital poplars tended to be larger than the senesced poplars, therefore, the vitality of the poplars should be taken into account when estimating the sapwood area.There were certain similarities and differences between the sap flow processes of vital and senesced poplars; sap flow velocity and diurnal water consumption of vital poplars was larger than that of the senesced poplars.However, the influence of environmental factors on sap flow was similar; sap flow of both vital and senesced poplars had a positive correlation with air temperature, solar radiation, vapor pressure deficit, groundwater depth, and negative correlation with air humidity.Our results demonstrated that the hydraulic characteristics of two different vitalities of poplar were not the same, and that when analyzing the water usage strategy and estimating the water consumption of poplar forests, these two kinds of trees should be treated differently.
and "Sino-German interdisciplinary joint training of innovative talents for ecological restoration of floodplain" funded by China Scholarship Council (Grant No. 201505990312).We thank the Forestry Department of Qarkilik (Ruoqiang), especially Askar and Sadik from Arghan Forestry Station for their support during our field work.We also would like to express our gratitude to Zhang Ximing (CAS), Atawulla Muyibul for supervising and supporting our sap flow equipment installation, and to our team memmbers Gulipiya Wumaier, Kailibinuer Nuermaimaiti and Maidina Muhetaer (Xinjiang University) for their assistance during the field work.Finally, we thank the editors and anonymous reviewers who helped to improve the manuscript.

Figure 1 .
Figure 1.The lower reaches of the Tarim River and the study area Arghan [10].

Figure 1 .
Figure 1.The lower reaches of the Tarim River and the study area Arghan [10].

Figure 2 .
Figure 2. Sapwood and inner wood of the tree stem where (a) the blue colored part is the sapwood, and the green colored part is the inner wood; (b) the white area on the left is the sapwood, the dark area on the right is inner wood of P. euphratica.

Figure 2 .
Figure 2. Sapwood and inner wood of the tree stem where (a) the blue colored part is the sapwood, and the green colored part is the inner wood; (b) the white area on the left is the sapwood, the dark area on the right is inner wood of P. euphratica.

Figure 3 .
Figure 3. Sapwood area estimation model for the vital poplars.

Figure 3 .
Figure 3. Sapwood area estimation model for the vital poplars.

Figure 4 .
Figure 4. Validation result of the sapwood area estimation model of vital poplars.

Figure 5 .
Figure 5. Validation result of the sapwood area estimation model.

Figure 4 . 14 Figure 4 .
Figure 4. Validation result of the sapwood area estimation model of vital poplars.

Figure 5 .
Figure 5. Validation result of the sapwood area estimation model.

Figure 4 .
Figure 4. Validation result of the sapwood area estimation model of vital poplars.

Figure 5 .
Figure 5. Validation result of the sapwood area estimation model.

Figure 9 .
Figure 9. Accumulated sap flow of vital and senesced poplar trees.

Figure 9 .
Figure 9. Accumulated sap flow of vital and senesced poplar trees.

Fv = − 50 .Figure 10 .
Figure 10.Diurnal variation of sap flow velocity of P. euphratica with different vitality and environmental factors.(A): Diurnal variation of sap flow velocity of different vitality poplars, (B): Diurnal variation of air temperature, (C): Diurnal variation of solar radiation, (D): Diurnal variation of air humidity, (E): Diurnal variation of vapor pressure deficit (VPD), (F): Diurnal variation of groundwater depth.

FFigure 10 .
Figure 10.Diurnal variation of sap flow velocity of P. euphratica with different vitality and environmental factors.(A): Diurnal variation of sap flow velocity of different vitality poplars, (B): Diurnal variation of air temperature, (C): Diurnal variation of solar radiation, (D): Diurnal variation of air humidity, (E): Diurnal variation of vapor pressure deficit (VPD), (F): Diurnal variation of groundwater depth.

Table 1 .
Summary statistics of poplar trees with vital and senesced growth status.

Table 1 .
Summary statistics of poplar trees with vital and senesced growth status.
Validation result of the sapwood area estimation model.

Table 2 .
Spearman correlation between sap flow velocity and environmental factors.

Table 3 .
Relationship models between sap flow velocity and environmental factors.

Table 2 .
Spearman correlation between sap flow velocity and environmental factors.