Study area

The Taklimakan desert is the largest desert in China, at the same time, is the world’s second-largest drifting desert, and covers an area of 33.7 square kilometers (^{Sun & Liu, 2006}). It is located in the center of Tarim basin in northwest China. This basin is a major area for producing oil and gas resources in China. The Tarim Desert Highway crosses the Taklimakan desert from south to north. The shelterbelt forests were built on both sides of the highway with the width of 72-78m. Due to the distance of the road across, the drifting desert is the longest in the world (^{Alles, 2012}), so it is listed in the Guinness Book of Records.

From north to south, Tarim Desert Highway sequentially crosses five major geomorphic units-compound transverse dune area, dome-shaped dune area, high compound longitudinal dune area, compound longitudinal dune area, and compound transverse dune area (^{Lei et al., 2008}). The study area is located in a dome-shaped dune area and the geomorphic unit has a total length of about 20 km, and the ground-water level is generally around 2 m. The water supply for the shelterbelt is basically ground-water with total dissolved solids (TDS) was 4.138 g/l and pH value is 7.21. The main pipe was used to convey the water from the ground while the branch pipe diverted the water with drippers installed on the drip tube for irrigation. The Aeolian sandy soil is fine sand, the soil bulk density of peripheral plantations is 1.43 g/cm³, and the soil porosity is 45.55%.

Test design

In order to effectively use the groundwater characteristics of the regions with relatively rich groundwater resources and to achieve water-saving irrigation, three additional water supply schemes for conservation of water resources were proposed based on the original irrigation system for shelterbelt project. As shown in figure 1, the entire shelterbelt land irrigated by well water was divided into four regions: region I with the existing irrigation system and regions II, III, and IV with modified irrigation systems for the comparison with region I (table 1).

Figure 1 Schematic diagram of the experimental layout. 

Table 1 Experimental design for different irrigation treatments. 

Since the Tarim Desert Highway Ecological Shelterbelt Project completion in 2005, the annual irrigation cycle from March to October was used for the shelterbelt. Under the condition of sufficient water resources for the spring and autumn irrigations, the above irrigation period was elongated, so that plants were under water stress compared to the original irrigation system. The growth of plants in different regions was monitored, in order to achieve effective water conservation without affecting plant growth. The specific experimental design is shown in table 1.

The data in table 1 are calculated as follows

Annual irrigation times (W_IT):

(1)

Irrigation water (W_IW):

(2)

Water conservation rate (W_CR). We set irrigation treatment I′ for W_IW = W_IWCK:

Then:

(3)

Here the number of days with irrigation is 240 from March to October. The water volume that drips out of emitters is 3.5 × 10^-3 m³ per hour. The control area of the drip water emitter is 1 m². The single irrigation time is 6 hours per emitter. W_IP is irrigation period/interval, W_{IP_1},W_{IP_2},W_{IP_3} are March to May, June to August, and September to October irrigation periods, respectively; n is II, III, IV.

Measurement and analysis

The effect that saline water irrigation to the groundwater was observed by four wells from the emitter every 30 cm called 1, 2, 3, 4, and set a control check (CK) observation well in bare sandy land around 100 meters away (figure 1). It reflected cumulative-change values by subtracting groundwater level and chemical indicators (TDS, pH, total alkalinity & total hardness) before the irrigation from the corresponding value after irrigation, where TDS were tested by using the 180 ºC drying weight method (^{APHA-AWWA-WEF, 1998}). Measurement of pH was performed using a pH meter (Model PHS-3C, Shanghai Lici Ltd., China). Total alkalinity was determined by titration with acid (^{Larocque et al., 2016}). Total hardness was determined by an EDTA complexometry.

The effect of the irrigation to the groundwater level was observed every day in an irrigation period 12d. It measured by releasing a probe connected with a multimeter by dual-cable wire into the well pipe. The multimeter was switched to the resistance shift. When the probe touched the water table, the cable was marked at the height of the well head, and the distance between the mark and the probe was measured after retrieval. The subtraction of the nth water table observation H_n data from the initial observation data H₀ gave the relative water level at the nth observation (H₀−H_n) to each observation wells.

Seedling investigation method was used for shelterbelts trees during the whole growing season, ten representative plants were randomly selected by human from Tamarixramosissima and Haloxylonammodendron respectively, and three indicators of plant height, crown width, and diameter at the ground level were measured. The One-Way ANOVA and least significant difference (LSD) method were used to analyze the influence of irrigation water for plant growth using SAS statistical software (version 9.1, SAS Institute Inc.).

Regular on-site investigation method was adopted to observe the soil salt in the experiment area on 1d, 6d, and 9d after irrigation. Soil electrical conductivity (EC) exhibiting a linear relationship with soil salinity within the study area (^{Wang, Xu, Lei, Li, & Wang, 2010}), the respective EC value was measured, in order to assess the soil salinity, by a SY-3 conductivity-thermometer (^{Institute of Soil Science, Chinese Academy of Sciences, 1987}) with the weight of water: soil = 5:1 (^{Han, Cao, Yimit, Xu, & Zhang, 2012}). The soil samples were collected at depths of 0, 5, 15, 25, 35, 55, 75 and 95 cm under the emitter with the horizontal distances of 0 to 150 cm every 30 cm from the emitter. Three samples per each fixed point were used to measure EC.

In each experimental area, a 2 m long Soil moisture observation tube was embedded into soil under emittter made of aluminum/plastic composite pipe of 50 mm in diameter. Soil moisture was determined via neutron probes (^{Gardner & Kirkham, 1952}) with five replicates after irrigation, then the mean value of replications was used as the soil water content. This rapid, in situ, nondestructive means of measuring soil moisture profiles in forestry (^{Wells & Fityus, 2011}) was used in this work to determine soil moisture at 10-cm intervals from the soil surface to a depth of 180 cm.

Spatial distribution of soil water and salt under different irrigation conditions

Water spatial distribution

Four different irrigation cycle experiments during irrigation season were carried out in four different regions of well irrigation area (table 1). According to figure 2, after the irrigation, the soil water content in region I with the original irrigation system reached saturation at a depth of 100 cm, while the soil water content in regions III, II, and IV reached saturation at depths of 160, 180, and 160 cm, respectively. Due to the shallow groundwater in the entire irrigation area, an aquifer is present within 180 cm soil depth, while the groundwater depths of four irrigated areas were different. The shallowest and the deepest groundwater depths were found in regions I and IV, respectively, with a gap of about 60 cm. Therefore, irrigation has a relatively high impact on the surface soil water, but a feeble one on deep soil water; the soil water content of each region can be maintained at a high level with the soil porosity of 45.55%.

Figure 2 Differences in the soil water content under different irrigation treatments. 

Salt temporal change

Since shelterbelt plants have high tolerance to soil salinity, plant could grow normally under various irrigation systems, and soil salinity increases from spring to autumn (^{Huang et al., 2015}). Due to multiple repeated irrigation cycles, salts dissolved in irrigation water may increase soil salinity, resulting in a salinity level higher than that tolerated by the plant (^{Zhang et al., 2008}), so understanding of the patterns of the salt migration and accumulation is very important. For example, the pattern in salt migration of region I was studied by continuous salinity monitoring within the irrigation cycle. As shown in figure 3(b), 3d after irrigation, the total soil salinity was significantly reduced, as compared to that in figure 3(a) and there was no significant salt accumulation. Most of the salt was infiltrated with the soil water from the range defined by the dashed line in figure 3(a) (i.e. the range from the surface to 20 cm) through the vadose zone into groundwater body. As can be seen from figure 3(c), over the course of evaporation and water redistribution, the salt gradually moved from the lower layer to the upper one in the vadose zone, and was accumulated in a depth range from 0~25 cm (dotted line) on day 9. Figure 3(d) shows that salt was continuously accumulated in the surface soil body at a horizontal distance of 20-140 cm from the dripper and a depth of about 20 cm.

Figure 3 Soil salt migration process within the irrigation cycle (a) Isoline of soil electrical conductivity one day after irrigation; (b) Isoline of soil electrical conductivity three days after irrigation; (c) Isoline of soil electrical conductivity six days after irrigation; (d) Isoline of soil electrical conductivity nine days after irrigation. 

Differences in plant growth under various irrigation volumes

With the survey method for shelterbelt tree seedlings, ten representative plants were randomly selected for the measurement of plant height, crown width, and diameter at the ground level. These characterization parameters are used to reflect the impact of different irrigation systems on plant growth. Single-factor analysis of variance was used to analyze the data from statistical experiments, and the respective results are presented in table 2.

Table 2 Variance analysis of the increment in plant growth with different irrigation treatments for selected plants. 

Note: * There is significant difference at p = 0.05; in the plant height, crown width, and diameter at the ground level.

The comparative analysis of the above results revealed no significant differences on all indicators of various treatments, except for plant height of Tamarixramosissima. This indicates that various irrigation systems have no significant effect on plant growth. The plant height difference for Tamarixramosissima can be assessed using the multiple comparison procedure among groups, which yields the results as shown in table 3. As can be seen, no significant difference was found between regions I and II, while there were differences found between region I and region III or region IV at the significance level of 5%. At a high significance level of 1%, there were differences between regions I and III.

Table 3 Analysis of height difference of Tamarixramosissima using the multiple comparison procedure (LSD). 

Impact on groundwater

Groundwater level

The values of the groundwater level changes at different observation wells during each irrigation period were obtained by subtracting the groundwater level after irrigation from the groundwater level before irrigation (figure 4). According to the changes in the groundwater level at different observation wells, the water levels in the four groundwater observation wells in the shelterbelt area showed an upward trend. The highest water level at No. 1 observation well was reached on day 10, and the highest water levels at the remaining three observation wells were reached on day 9. On the other hand, no obvious change pattern was found in the groundwater level in the observation well CK that was located in the original sands outside of the shelterbelt. According to observational data, the largest and the smallest rises of water levels were 0.022 m (No. 3 observation well) and 0.001 m (No. 1 observation well), respectively, indicating that irrigation can raise groundwater levels, but this effect is feeble, resulting only in few centimeter changes.

Figure 4 Changes in groundwater level at different observation wells in the shelterbelt area. 

Groundwater chemistry

In this study, groundwater chemistry was indicated by pH, TDS, total hardness and total alkalinity (^{Salem, Atwia, & El-Horiny, 2015}). The value of cumulative-change as relative comparison was carried out to obtain the change values compared with the initial value, such as heavy metal accumulation (^{Unterbrunner et al., 2007}), plant relative growth yield (^{Keyantash, & Dracup, 2002}), cumulative precipitation (^{Alkhayri & Albahrany, 2004}), etc. Figure 5 shows that the main characteristic changes in groundwater chemistry were reflected on the changes in TDS and total hardness. The changes in the above mentioned two indicators decreased within the shelterbelt area, while the changes of the same two values increased for the bare sandy land outside of shelterbelt land (the change was positive). Therefore, under the influence of irrigation projects, the groundwater environment in the local areas allowed the TDS and total hardness of groundwater in the shelterbelt area covered by the irrigation system to exhibit the significant upward trend. However, values of pH and total alkalinity showed no significant changes.

Figure 5 Cumulative changes in each indicator of water chemistry at different observation wells in the shelterbelts area.