Water-saving Potential in aeolian sand soil under straight tube and surface drip irrigation in Taklimakan Desert in Northwest China

ZhongWen Bao , HuLin Du , XiaoJun Jin

1. Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Division of Hydrology and Water-Land Resources in Cold and Arid Region, Lanzhou, Gansu 730000, China 2. Graduate University of Chinese Academy of Sciences, Beijing 100049, China 3. Tarim Oilfield Corporation, China National Petroleum Corporation, Korla, Xinjiang 841000, China

Water-saving Potential in aeolian sand soil under straight tube and surface drip irrigation in Taklimakan Desert in Northwest China

ZhongWen Bao1,2, HuLin Du1, XiaoJun Jin3

1. Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Division of Hydrology and Water-Land Resources in Cold and Arid Region, Lanzhou, Gansu 730000, China 2. Graduate University of Chinese Academy of Sciences, Beijing 100049, China 3. Tarim Oilfield Corporation, China National Petroleum Corporation, Korla, Xinjiang 841000, China

Evaporation loss from the saturated soil beneath drip irrigation emitters highly influences the irrigation efficiency of drip irrigation(DI). Subsurface drip irrigation (SDI) is one good approach to curb this inefficiency, but in a new irrigation method, straight tube irrigation (STI), the irrigation tubes do not need to be buried and thus STI is recommended to increase the irrigation efficiency under normal surface-applied DI. STI consists of only connectors and water-transference tubes that can directly transfer irrigation water from the lateral emitters in the drip line to the root zone of plants. Five-month field experiments were carried out in aeolian sand soil in the forest-belts of the Taklimakan Desert, which have poor water storage capacity, to compare the potential water saving between STI and DI. The preliminary results showed that, compared with DI, STI (1) improved the soil water content in soil depths from 40 to 100 cm under the soil surface; (2) achieved the same irrigation effects in relatively shorter irrigation durations;(3) had very little water loss due to deep seepage; and (4) formed a layer of dry sand about 10 to 30 cm thick immediately below the soil surface, which lessened evaporation loss of soil water beneath the emitters on the soil surface. This demonstrates that STI can maximize the water-saving potential of DI through the reduction of wetted soil perimeters on the soil surface. This is valuable information for water-saving engineering applications and projects with STI in arid and semiarid regions.

potential water saving; evaporation loss; straight tube irrigation; drip irrigation; efficiency; Taklimakan Desert

1. Introduction

Water availability is one of the most decisive factors in agricultural production, economical construction, residential life, ecological restoration, and environmental protection in arid and semiarid regions in northwest China. Thompsonet al.(2009) pointed out that the combination of increasing urban population, increasing drought areas, declining groundwater levels, and loss of water storage in reservoirs is putting water resources in arid and semiarid regions at severe risk. To address the competing needs for water in industrial and residential construction, agricultural production,and environmental protection, effective irrigation methods such as border irrigation (BI), sprinkler irrigation (SI), surface drip irrigation (DI), subsurface drip irrigation (SDI),and sand tube irrigation (Meshkatet al., 1999a) are gradually being adopted to improve irrigation efficiency in such areas. Kandelous and Simunek (2010) concluded that DI and SDI systems are the most frequently used water-saving systems for irrigating crops, vegetables, and forest-belts in arid and semiarid regions.

However, DI and SDI systems have several disadvan-tages that need to be overcome during their application. For example, in SDI the laterals must be buried and the underground emitters often clog (Blass 1971; Camp, 1998; Dasberg and Or, 1999; Campetal., 2000); this has limited the application of SDI in regions with limited water resources.With DI, the evaporation loss from the constantly saturated soil (i.e., the soil wetted perimeters beneath the emitters in the DI drip lines) can lead to poor water management and lack of water resources conservation. Xuet al. (2008) contended that there was still some water-saving potential under the DI system in the protection of the forest-belts in the Taklimakan Desert, but Meshakt (1999b) concluded that DI results in less actual soil wetting, thus requiring more frequent irrigation which would cause the soil surface to remain wet and the first stage of evaporation to persist longer,resulting in unacceptable rates of water loss. The efficiency of DI is highly influenced by the evaporation losses from the soil surface wetted perimeters.

The deep drainage of irrigation water is another important factor influencing the water available for plant use.Therefore, it became necessary to develop a new irrigation method that is easy to install and apply and can effectively lessen the evaporation loss and deep drainage from the wetted perimeters under DI.

Elmaloglou and Diamantopouls (2009) stated that DI using buried emitters (SDI) has the potential to save irrigation water by reducing the soil surface wetted perimeters,which would reduce evaporation loss of soil water from the soil surface. Therefore, SDI could be applied to enhance agriculture and fruit orchard productivity, increasing water use efficiency while almost completely eliminating evaporation loss of soil water from the soil surface (Phene, 1995;Camp, 1998; Campet al., 2000; Patel and Rajput, 2008;Thompson, 2009). If the difficulty of installation and the clogging of the underground emitters were properly overcome, SDI might be widely applied to reduce evaporation loss of soil water content.

Meshkat (1997) and Meshkatet al. (1998, 1999a,b) devised an irrigation method called "sand tube irrigation" and conducted laboratory experiments on evaporation and moisture distribution using undisturbed and reconstructed soils.Comparisons of sand tube irrigation versus DI in laboratory conditions have indicted that sand tube irrigation significantly reduced evaporation in undisturbed soil monoliths.However, sand tube irrigation is appropriate only for permanent tree crops because it requires the removal of a soil core beneath each emitter and requires filling the voids with coarse sand. In addition, in the Taklimakan Desert in northwest China, sand tube irrigation is not feasible in the aeolian sand soil due to the soil characteristics and its narrow limitations of application.

In this study we investigated a new form of irrigation,straight tube irrigation (STI), which employs a normal applied surface drip system in conjunction with water-transference tubes (straight tubes) and connectors. The entire system is designed to eliminate the need to bury any components, and to significantly reduce evaporation loss.STI is actually a modified SDI method. The objective of this research was to investigate the potential water saving under STI through field experiments and to clearly describe the structure, underlying principles, and installation methods of STI. Included are comparisons of the moisture infiltration and redistribution in the soil profile under the same irrigation durations with various sampling intervals under DI and STI. The field experiments were performed in aeolian sand soil in the forest-belts along the Tarim Desert Highway in the Taklimakan Desert for 5 months in 2008 from May to September.

2. Materials and methods

2.1. Climate conditions and soil characteristics of the ex-

perimental site

The field experiments were carried out at a site near the No.49 irrigation motor-pumped well in the cross-hinterland forest-belts of the Taklimakan Desert (39°06′N, 83°40′E).This area has an annual mean temperature of 12.4 °C; the annual accumulated temperature above 10 °C is 4,261.8 °C;the annual precipitation is 11.05 mm and its distribution is irregular; the annual average relative humidity is 29.4%; the potential evaporation is 3,638.6 mm; the annual sunshine duration are 2,571.3 hrs; and the average wind velocity is 2.5 m/s (Huang, 2002; Liet al., 2005; Zhouet al., 2006; Xuet al., 2008). The soil is predominantly aeolian sand. According to a particle size analysis of the soil samples, the aeolian sand soil is primarily composed of extremely fine sand (0.05–0.10 mm) and normal sand (0.10–0.25 mm),which accounts for 67%–91% of the soil.

The forest-belts along the Tarim Desert Highway were established in 1995. The main selected species of plants to reforest the desert highway and the base of the Tarim oilfield wereHaloxylon ammodendron,Tarmarix ramosissima, andCalligonum arborescens.They were intermixed and planted in rows spaced 1 m apart; the spacing in the rows was 1 m.The irrigation method used in the forest-belts is currently DI and the irrigation interval is 8–12 d. The depth of the underground water table is 4–10 m, the mineral content of the water is 2.26–31.87 g/L, and its pH is 8.13 (Zhouet al.,2000, 2006; Xuet al., 2008).

2.2. Introduction of straight tube irrigation

STI is based on the DI system and is a modified SDI system. DI is defined as the application of water through point or line sources (emitters) on or below the soil surface at a low operating pressure (20–200 kPa) and at a low discharge rate (1–30 L/hr per emitter), resulting in partial wetting of the soil surface (Dasberg and Or, 1999). SDI is defined as the application of water below the soil surface through emitters, with discharge rates generally in the same range as DI. Thus with DI there is significant evaporation loss of soil water content from the soil surface wetted perimeters under the emitters. SDI, although having the disadvantages of difficult installation and the potential for clogging, can effectively lessen the evaporation loss of soil water through the reduction of ponding areas on the soil surface.The new irrigation method, STI, is shown in Figure 1. It is easy to install and maintain (like DI) and, if effectively applied to lessen the evaporation loss of soil water from the soil surface (like SDI), it can be adopted to combine the advantages of SDI and DI.

Figure 1 The structure and installation of STI: (a) the general structure of STI with a connector and a water-transference tube;(b) the connector that is used to connect the lateral emitterin the drip line and the water-transference tube;(c) STI installed in the field site.

A connector and a water-transference tube are the primary components of STI. The water-transference tubes,which directly transfer water from the emitters above the soil surface to the root zone of the plants, ensure water infiltration from the emitters to the root zone of the plants by the shortest possible route.

In STI, the design of the water-transference tubes must consider the depth of the root distribution of the plants in order to determine the best length of the tubes. A funnel in the bottom of each tube prevents clogging during the inserting process. The tubes have 1.2-mm-diameter micro pores to ensure that water infiltrates into the root zone of the plants.The purpose of the connector is to connect the emitter and the water-transference tube. The design of the connector is related to the size of the drip tube and the water-transference tube. In aeolian sand soil regions, two micro baffles are built into the connector to prevent fine sand from entering it. With these two components, STI can achieve the goal of transferring water from the soil surface directly to the root zone of the plants to enable root zone irrigation.

2.3. Experimental design and measurements

The DI system extracts high-mineral groundwater from motor-pumped wells to irrigate plants; this method is widely applied for eco-environmental protection in the base of the Tarim oilfield and the major highway in the Taklimakan Desert. Under DI, the irrigation water slowly infiltrates from the lateral emitters on the soil surface, through the soil profile, to the root zone of the plants. In this study, components of STI and DI were alternately installed in the forest-belts according to the lateral emitters in the drip line (nothing need to be installed for DI), as shown in Figure 2. The irrigation flux was designed to apply water at a steady rate of 3.6 L/hr. During one period the irrigation duration was 6 h and the volume of irrigating water was approximately 21.6 L per emitter. Field analyses of the vertical distribution of soil water content under STI and DI were then conducted. According to a previous survey of the distribution of roots of the experimental plants in the forest-belts, the length of the STI water-transference tubes was designed to be 40 cm.

The field experiments were divided into four groups and carried out from May to September in 2008. The first group of experiments were initialized to probe the horizontal distribution of soil water content after 6-hr irrigation durations under STI in May. Soil sampling was conducted in 24-hr intervals at distances of 10, 30, and 50 cm from the center of a lateral emitter. The second group of experiments investigated the vertical distribution of soil water content at different sampling intervals in August. More than 200 plants, half of which were irrigated under STI and the rest under DI,were irrigated for 6 hrs. The intervals at which to measure the soil water content after the 6-hr irrigation duration were set to 10 hrs, 12 hrs, 36 hrs, 60 hrs, and 6 days. The third group of experiments addressed the vertical distribution of soil water content after different irrigation durations under STI and DI in September. These experiments were divided into three segments which applied water at a steady rate of 3.6 L/hr for irrigation durations of 3-hr, 4-hr, and 5-hr. Soil samples were collected immediately after each irrigation.The fourth group of experiments in September had 5-hr irrigation durations and sampling intervals of 21-hr.

The overall investigation also compared the effects of different lengths of STI water-transference tubes. The lengths of the water-transference tubes were set to 0 cm (DI),and 20, 30, and 40 cm for STI. Also, the thickness of the layer of dry sand under STI was measured during the field experiments.

Soil samples were taken by pressing a 100-cm-long,5-cm-inside-diameter steel soil sampling tube horizontally into the soil profile at selected coordinate positions at a distance of 10 cm from the STI and DI emitters. The total sampling depth was 200 cm and the soil sampling layer increment was set to 20 cm (unless specifically mentioned in this paper, the sampling distance from the lateral emitters was 10 cm and the length of the STI water-transference tube was 40 cm). The soil water content was measured by gravimetric moisture, which was determined by calculating the proportion of water loss relative to dry soil weight after oven-drying the soil samples.

Figure 2 Schematic of the experimental design under DI and STI

3. Results and discussion

3.1. Comparison of STI and DI vertical distributions of soil water content after 6-h irrigation duration with different sampling intervals

Figure 3 shows the vertical distributions of DI and STI soil water content in the same drip line after a 6-hr irrigation duration with five different sampling intervals. The sample intervals were set to 10 hrs, 12 hrs, 36 hrs, 60 hrs, and 6 days.Figure 3a illustrates that the vertical distribution of soil water content under DI could be divided into four different water-storage layers. The first layer, called the "quickly changing layer", was located 20 cm under the soil surface. Due to the severe influence of the outside environment and intense evaporation, the relative humidity of the surface soil profile evidently varied considerably. According to site measurements, the soil water content of this first layer was saturated right after DI; however, after six or more days there was little soil water content remaining in the surface soil profile.The second layer, called the "median changing layer", was located 20–60 cm under the soil surface. Because of infiltration and redistribution of soil water content under DI, this layer incurred the most active soil water content changes.The third layer, called the "poorly changing layer", was located 60–120 cm under the soil surface. The soil water content in this water-storage layer changed slightly as a result of infiltration and redistribution of soil water content and root water uptake function. The fourth layer, called the "stable layer", was located more than 120 cm under the soil surface.Due to the subtle impact from DI and groundwater, the soil water content in this water-storage layer varied only slightly with the soil depth.

Figure 3b illustrates the vertical distribution of soil water content under STI, demonstrating that irrigation water was directly transferred to the depth of 40 cm under the soil surface without any influence on the first water-storage layer(the dry sand layer). According to the site samples, there existed a 10- to 30-cm layer of dry sand that could effectively lessen the evaporation loss of soil water content from the soil surface. Thus, this dry sand layer functioned as an"umbrella" that prevented evaporation loss of irrigation water. In the second water-storage layer namely "first rapid changing layer", located 20–40 cm under the soil surface, the distribution of soil water content was most affected by the STI irrigation water and the sandy soil character. The third water-storage layer namely "second rapid changing layer", located 40–100 cm under the soil surface, was less affected than the second layer but the soil water content in both of these layers responded very differently (oppositely) under STI compared to DI. The fourth layer also namely "stable layer",located below 100 cm, was negligibly affected by irrigation water, the outside environment, or the groundwater.

These data suggest that STI can effectively reduce water evaporation on the soil surface and improve the soil water content at soil depths from 40 to 120 cm. Moreover, the vertical distribution of soil water content of sample No. 99(under STI) and sample No. 102 (under DI) was considered to be the "stable" soil water content distribution after the 6-h irrigation duration in the 6-d sample interval. The curve of the soil water content of sample No. 99 illustrates an optimal vertical distribution of soil water content that can better irrigate plants in a sandy region and significantly lessen the evaporation of soil water from the soil surface. The curve of sample No. 102 illustrates that DI cannot avoid the evaporation of soil water near the emitters in the drip line, and that the highest amount of soil water occurred 20 cm higher in the soil profile than where the majority of the plant roots were located, which is of very little benefit to the growth of the plants in the forest-belts.

Figure 3 Vertical distribution of the soil water content varied with soil depth: (a) DI; (b) STI.

3.2. Comparison of distribution of soil water content of different irrigation durations under STI and DI

To determine the vertical distribution of soil water content after irrigation durations of 3-hr, 4-hr, and 5-hr, three segments of experiments under STI and DI were carried out in the same drip line. The soil samples were collected immediately after each irrigation duration.

Figures 4a, c, and e illustrate that STI did not change the distribution trend of soil water content in the vertical direction except to cause some increase of soil water content at the depth from 20 to 100 cm. Figures 4b, d, and f show that the range of vertical distribution of soil water content under STI at the depth from 30 to 100 cm evidently increased with the persistence of irrigation. As depicted in Figures 4a and b,after the 3-hr duration the average soil water content after STI was 2.88%, which was 0.05 times more than that before STI; the average soil water content after DI (3.13%) was 0.2 times more than that before DI. Conversely, as illustrated in Figures 4c and d, after the 4-h duration the average soil water content after STI was 2.83%, which was 0.5 times more than that before STI; the average soil water content after DI(2.83%) was only 0.14 times more than that before DI.Similarly, Figures 4e and f show that after the 5-hr duration the average soil water content after STI was 3.69%, which was 0.45 times more than that before STI; the average soil water content after DI (2.59%) was only 0.15 times more than that before DI.

This suggests that STI and DI could both be utilized to improve the soil water content, but the increasing range of soil water content under STI was much larger than that under DI after 4-hr and 5-hr durations. In addition, the relative proportions of the average soil water content between STI and DI were 0.92:1 after the 3-hr duration, 1.15:1 after the 4-hr duration, and 1.42:1 after the 5-hr duration. This indicates that under the same soil profile, climate, water resources, irrigation duration, and other related environmental elements, STI could improve the irrigation efficiency much more than DI.

3.2.1 Comparison analysis of the vertical distribution of soil water content between STI and DI

In the vertical distribution of soil water content under STI discussed above, water could be directly transferred to the root zone of the plants through the water-transference tube. Theoretically, STI could more significantly reduce water loss compared to DI, especially in arid and semiarid regions that experience severe evaporation. Therefore, in order to compare irrigation efficiency between STI and DI,we compared the average soil water content between a relatively shorter irrigation duration under STI and a relatively longer irrigation duration under DI. Figure 5a illustrates the curves of the vertical distribution of soil water content in a 3-hr duration under STI and a 4-hr duration under DI. The average soil water content of the 3-hr duration under STI(2.88%) was 0.17 times more than that of the 4-h duration under DI at the depth from 0 to 200 cm. That of the 3-hr duration under STI (3.36%) was 0.08 times more than that of the 4-hr duration under DI at the depth from 0 to 100 cm.In addition, the maximum soil water content under STI(10.01%) was 0.59 times more than that under DI. Similarly,in Figure 5b, the average soil water content of the 4-h duration under STI (2.83%) was 0.09 times more than that of the 5-hr duration under DI at the depth from 0 to 200 cm. That of the 4-hr duration under STI (3.57%) was 0.12 times more than that of the 5-hr duration under DI at the depth from 0 to 100 cm. Hence, relatively shorter irrigation duration under STI could improve the soil water content, which would need a longer duration under DI. If STI were widely applied in the forest-belts, its relatively shorter irrigation duration requirements could reduce groundwater use, evaporation loss,and energy use.

Figure 4 Vertical distribution of soil water content with different irrigation durations varied with soil depth: (a) 3-hr irrigation duration:comparison of soil water content between before STI and after STI; (b) 3-hr irrigation duration: comparison of soil water content between after STI and after DI; (c) 4-hr irrigation duration: comparison of soil water content between before STI and after STI; (d) 4-hr irrigation duration: comparison of soil water content between after STI and after DI; (e) 5-hr irrigation duration: comparison of soil water content between before STI and after STI; (f) 5-hr irrigation duration: comparison of soil water content between after STI and after DI.

Figure 5 Comparison analysis of vertical distribution of soil water content between different irrigation durations varied with soil depth.(a) comparison between a 3-hr duration under STI and a 4-hr duration under DI;(b) comparison between a 4-hr duration under STI and a 5-hr duration under DI.

3.2.2 Horizontal distribution of soil water content under STI

Both STI and DI systems produce wetted perimeters under the emitters. As depicted in Figure 6, three curves of vertical distribution of soil water content varied with soil depths at 10, 30, and 50 cm from the center of an STI emitter,observed in field experiments and indoor analysis. The soil water content greatly varied at depths from 0 to 80 cm under the soil surface, but remained essentially unchanged below the depth of 80 cm. This indicates that STI not only improves the soil water content in the layer where the majority of the distribution of plant roots exist (the root zone), but also can effectively restrain the deep percolation of irrigation water.

Figure 6 Horizontal distribution of soil water content varied with soil depth

3.2.3 Vertical distribution of soil water content under STI with different lengths of the water-transference tube

Figure 7 shows the vertical distribution of soil water content under STI with different lengths of the water-transference tube: 0 cm (DI), 20 cm, 30 cm, and 40 cm.In general, the soil water content produced by the 20-cm-long water-transference tube varied less than that produced by DI, but the curves of vertical distribution of soil water content under STI with 20-cm-long water-transference tube and DI are contrary to those under STI with 30-cm-long and 40-cm-long water-transference tubes. The data indicate that only the water-transference tubes more than 30 cm in length can produce effective root-zone irrigation. In addition,the soil water content with the 30-cm-long water-transference tube at the depth of 0–40 cm under the soil surface is much greater than that of the 40-cm-long water-transference tube, which suggests that STI with a 30-cm-long water-transference tube would cause much more evaporation of soil water from the soil surface. Therefore,the 40-cm length would save much more water in the forest-belts, which is in accord with the survey of the root distribution of the experimental plants.

4. The impact of STI on soil water loss through evaporation

Under STI, water can be directly transferred to the soil layer at the depth of 30–40 cm under the soil surface, lessening water loss due to evaporation from the soil surface. A batch of emitters of STI with 40-cm-long water-transference tubes was installed in July. Sampling 20 days later indicated an 18- to 20-cm-thick layer of dry sand under most of the STI emitters; 30 days later there was a 30-cm-thick layer of dry sand under several of the STI emitters. Field experiment results showed the soil water content of dry sand at the depth of 10 cm under the soil surface was 0.13%–0.15%, which was similar to the results of Wanget al. (2008), where the soil water content on the soil surface layer of moving sand dunes was 0.1%–0.3%. Also, Liuet al. (2006) found that if there existed a dry sand layer more than 30 cm thick, there would be 92.38% less water loss from soil evaporation of irrigation water. Hence, STI has the potential to lessen the evaporation loss of soil water from the soil surface without apparently impacting the surfacing issues. This has a positive impact on soil water content conservation, improves irrigation efficiency, and decreases evaporation water loss in sandy regions.

Figure 7 Vertical distribution of soil water content with different lengths of the water-transference tube under STI: 0 cm (DI),20 cm, 30 cm, 40 cm varied with soil depth.

5. Preliminary conclusions

Effective irrigation methods such as DI and SDI have the potential to save water while enhancing the productivity of crops and protecting the ecology and environment. However, even though SDI reduces the soil surface wetted perimeters beneath DI emitters, SDI is relatively difficult and expensive to install and has the potential for clogging of the emitters. These disadvantages hinder the wide application of SDI. STI, with its connectors and water-transference tubes,combines the advantages of SDI and DI and eliminates their disadvantages.

In 5-month field experiments, the vertical distribution of soil water content at various soil depths under STI and DI was analyzed to compare the potential water savings of each system. DI applied irrigation water to the soil surface and the water slowly infiltrated to the root zone, while STI directly transferred the water to the root zone. Analysis of the field data indicated that STI improved the soil water content in the root zone soil layer, while DI improved the soil water content in the soil surface layer. However, the high winds in the desert caused severe evaporation loss of irrigation water from the surface soil profile. In comparison, STI improved the soil water content at the depth of 40–100 cm under the soil surface and also created a layer of dry sand about 10–30 cm thick under the soil surface; this dry sand soil profile had the potential to lessen the evaporation loss of soil water.Similar to DI, STI did not significantly affect deep water seepage. In other words, STI can theoretically directly transfer irrigation water to the root zone, which benefits the water uptake of the roots and the growth of the plants without irrigating the soil surface beneath the emitters (as in DI).This would limit the evaporation loss of soil water content on the soil surface. Also, STI formed a dry sand soil layer that acts as a protection cover to restrain the evaporation loss of water and effectively lessen water loss due to deep seepage. Thus, in many respects STI can improve the efficiency of irrigation and make the most of the water-saving potential of DI.

In conclusion, STI is more efficient than DI and the STI equipment is easy to install, maintain, and dismantle.Therefore, STI can play a crucial role in the protection of forest-belts in arid and semiarid regions and can be widely applied in the construction of forest-belts, ecological projects,and other water-saving practices in arid, semiarid, and desert zones. Further research on the dynamics of soil water content, saline water utilization, and effective irrigation schedules under STI will be conducted in the near future.

The authors wish to thank Tarim Oilfield Corporation of China National Petroleum Corporation for providing funds and the site for the experiments. The authors express their gratitude to the Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences for their support. Special thanks are given to Allan Grey (the English writing teacher at Chinese Academy of Sciences in Lanzhou in China) for correcting this manuscript, and the advice and suggestions of the editor and the reviewers.

Blass S, 1971. Drip irrigation. In: Drip (trickle) and automated irrigation in Israel. Water Commissioners Office, Ministry of Agriculture, Tel Aviv,Israel, 1: 10–28; cited by Dasberg S, Or D, 1999, Applied agriculture:Drip irrigation. Springer-Verlag, New York.

Camp CR, 1998. Subsurface drip irrigation: A review. Transactions of the ASAE, 41(5): 1353–1367.

Camp CR, Lamm FR, Evans RG, Phene CJ, 2000. Subsurface drip irrigation:Past, present, and future. In: Proc. 4th Decennial National Irrigation Symposium, 391–400. ASAE, St. Joseph, Mich.

Dasberg S, Or D, 1999. Applied Agriculture: Drip Irrigation.Springer-Verlag, New York, NY.

Elmaloglou S, Diamantopouls E, 2009. Simulation of soil water dynamics under subsurface drip irrigation from line sources. Agricultural Water Management, 96: 1587–1595.

Huang Q, 2002. Water and salt movement in soil of Taklimakan desert irrigated with saline water. PhD dissertation. Northwest University of Farming and Forestry Technology, Shanxi, China.

Li H, Li S, Lei J, Li B, Xu X, Zhou H, 2005. Experimental study on a forest station irrigated with saltwater with different mineralizations in the Taklimakan desert. Arid Land Geography, 28(3): 301–310.

Liu X, Zhang T, Zhao H, He Y, Yun J, Li Y, 2006. Influence of dry land bed thickness on soil water content evaporation in mobile dune. Arid Land Geography, 29(4): 523–526.

Kandelous MM, Simunek J, 2010. Comparison of numerical, analytical, and empirical models to estimate wetting patterns for surface and subsurface drip irrigation. Irrigation Science, 28: 435–444.

Meshkat M, 1997. Modeling and laboratory evaluation of evaporation and moisture redistribution for drip/trickle and sand tube irrigation using undisturbed and reconstructed soils. PhD dissertation. Biosystems and Agricultural Engineering Dept., University of Kentucky, Lexington, Ky.

Meshkat M, Warner RC, Workman SR, 1998. Comparisons of water and temperature distribution profiles under sand tube irrigation. Transactions of ASAE, 41(6): 1657–1663.

Meshkat M, Warner RC, Workman SR, 1999a. Evaporation reduction potential in an undisturbed soil irrigated with surface drip and sand tube irrigation. Transactions of ASAE, 43(1): 79–86.

Meshkat M, Warner RC, Workman SR, 1999b. Modeling of evaporation reduction in drip irrigation. Journal of Irrigation and Drainage Engineering ASCE, 125(6): 315–323.

Patel N, Rajput TBS, 2008. Dynamics and modeling of soil water under subsurface drip irrigated onion. Agricultural Water Management, 95:1335–1349.

Phene CJ, 1995. The sustainability and potential of subsurface drip irrigation.In: Freddie Lamm, (Ed.), Microirrigation for a changing world. ASAE,St. Joseph Mich., 359–367.

Thompson TL, Pang H, Li Y, 2009. The potential contribution of subsurface drip irrigation to water-saving agriculture in the western USA. Agricultural Science in China, 8(7): 850–854.

Wang Y, Xiao D, Li Y, 2008. Spatial and temporal dynamics of oasis soil salinization in upper and middle reaches of Sagonghe River, Northwest China. Journal of Desert Research, 28(3): 478–484.

Xu H, Zhang X, Yan H, Liang S, Shan L, 2008. Plants water status of the shelterbelt along the Tarim Desert Highway. Chinese Science Bulletin,53(Supp II): 146–155.

Zhou Z, Xu X, Lei J, Li S, 2006. Ecological stability of Tarim Desert Highway shelterbelt. Chinese Science Bulletin, 51(Supp I): 153–160.

Zhou Z, Xu X, Li B, 2000. Study on trends of water and salt of artificial greenbelt in hinterland of Takilimakan Desert. Arid Zone Research,17(1): 21–26.

10.3724/SP.J.1226.2011.00243

*Correspondence to: ZhongWen Bao, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Department of Hydrology and Water-Land Resources in Cold and Arid Regions, Lanzhou, Gansu 730000,China.Email: baozhwen@gmail.com

2 January 2011 Accepted: 16 March 2011