Assessing the impact of shallow subsurface pipe drainage on soil salinity and crop yield in arid zone

Study area

The study area is situated at Shihutan Township (85°52″–86°12″E, 44°31″–44°46″N, Fig. 1), Shihezi Oasis, Xinjiang Uyghur Autonomous Region, China. Covering an area of 20 ha, the experimental site is located in south margin of the Gurbantunggut Desert where severely salinized farmlands have been abandoned for 14 years. The annual average rainfall is 175 mm (1956–2010). The annual average temperature is 6.8 °C. The annual average evapotranspiration (ET) is 1200 mm (1956–2010). The groundwater table varies between 1.5 m to 3.5 m during the year in response to irrigation and evapotranspiration. The mineralization degree of groundwater is 5-7 g L⁻¹. The average mass fraction of soil organic matter is 11.4 g kg⁻¹, and the mean of soil pH value is 8.28. Secondary soil salinity decreases from east to west, with the highest reaching over 5 g kg⁻¹ and the lowest about 1.5 g kg⁻¹. Soil physical parameters at different depths are shown in Table 1. The crop type was cotton. The groundwater level at the midpoint of the distance between drainage pipes is 1.5 ∼2 m high in spring, and 2.5 ∼3.5 m low in summer and autumn.

Field experiment

Implementation of subsurface drainage systems

Subsurface pipes were laid on the salinized farmlands in Autumn of 2009. The drainage in soils was digged using the trencher which had a great power source and special cutting knives or bits. A drain depth in agricultural practice is usually recommended as >1.2 m to avoid damaging the laterals (Tiwari & Goel, 2015). In this study, drain depth was set at 2 m underneath the frozen soil layer in Shihezi. Drain spacing was taken as 80 m. A schematic diagram of field experiment set-up is shown in Fig. 2. Corrugated pipes made of polyvinyl chloride (PVC) have an inner diameter of 11 cm. They as drain pipes were coated by gauze and then backfill and compaction. Numerous soil and gravel in different sizes were used for backfilling. Size of soil and gravel gradually decreases from bottom to top. The drainage water in the collecting well was pumped to the shelter forest.

Surface experimental design

The experimental cotton field laid with subsurface pipes was divided into a mildly salinized plot with a salt content of 1–3 dS m⁻¹ and a moderate salinized plot with a salt content of 3-6 dS m⁻¹. Each covers an area of 8 ha (400 m in length, 200 m in width). The mildly and moderately salinized soils without drainage pipes were used as two control check (CK) plots. Barrier was installed to prevent water movement between the experimental plots and the CK plots. One plastic film, two drip irrigation tapes and six rows of cotton were arranged for drip irrigation under mulch (Fig. 3). The film width was 2.1 m, and the cotton crop was irrigated 11 times during the whole growth stages. The total irrigation quota was 4.5 million L ha⁻¹. Electrical conductivity (EC) of irrigation water was 0.35 dS m⁻¹, and was assumed to have no impact on soil salinity. Rainfall and potential evapotranspiration (ET₀) in the whole year are shown in Fig. 4. In this study, the total amount of nitrogen, phosphorus and potassium applied per acre of land during the growth of cotton is: nitrogen: 16.75 kg; phosphorus pentoxide: 2.76 kg; potassium oxide: 14.52 kg.

Sample collection, measurement and analysis

(1) Sample collection: soil samples were collected from depths of 0–20 cm, 20–40 cm, 40–60 cm and 60–80 cm at the four primary growth stages, seeding, budding, flowering and bolling. There were three replications for mild salinization soil, moderate salinization soil and the CK plots. Drainage water in subsurface pipes was collected from a collection well using a manual water sampler. Soil electrical conductivity (EC_s), electrical conductivity (EC_w) and the pH value of drainage water were measured. Diagonal method was employed to select three sampling points to collect plant samples before the harvest of cotton on August 20. Table 2 shows the number of plants, number of bolls and mass of bolls were measured for calculating the cotton yield (kg ha⁻¹). (2) Measurement: the EC of soil and drainage water in the subsurface pipes were measured by using an electrical conductivity meter (DDSJ-308A conductivity meter, water-to-soil ratio 5:1, dS m-1). The pH value of the drainage water was measured by Ampholine (LKB Producter AB, Sweden) pH meter (3310, water-to-soil ratio 2.5:1). Soil desalinization rate was calculated according to Eq. (1): (1)${\displaystyle D_r=\frac{S{S}_i-S{S}_f}{S{S}_i}\times 100\text{\%}}$

where D_r is desalinization rate of soil, SS_i is initial soil salinity, SS_f is final soil salinity.

Variation rate of drainage water salinity was calculated according to Eq. (2): (2)${\displaystyle V{R}_{dw}=\frac{E{C}_{dw}-E{C}_{idw}}{E{C}_{idw}}\times 100\text{\%}}$

where VR_dw is variation rate of drainage water salinity, EC_dw is EC of drainage water, EC_idw is initial EC of drainage water.

Cotton yield was calculated according to Eq. (3): (3)${\displaystyle CY=\frac{N_p\times N_b\times W_{sb}}{1000}}$

where CY is seed cotton yield (kg ha⁻¹), N_p is number of plants (plant ha⁻¹), N_b is number of bolls (bolls plant⁻¹), W_sb is mean weight of a single boll which is harvested randomly from different plants in rows of three sampling points (g boll⁻¹).

Statistical analysis

Statistical analysis was conducted using the statistical software SPSS version 20.0 (IBM, Armonk, NY). Data were reported as the mean ± standard error. Least significance difference (l.s.d.) was adopted for a variance analysis (Gong et al., 2015). The letter-marking method was used in the comparison of different treatments. Significant differences in ECs among different soil depths were indicated using different lowercase letters (P < 0.05). Significant differences in cotton yield between the treatments of subsurface pipe drainage and CK were indicated using different capital letters (P < 0.01) (Yang, Chen & Zhang, 2019).

Variation of soil salinity

Salinity of the moderately salinized soil decreased during all growth stages (Fig. 5). The decrease of ECs at the 0–20 cm depth was greater than those at other depths. Typically, the salt leaching effect was achieved through drip irrigation onto surface soil. After several applications of drip irrigation, soil salts migrated downwards with the wetting front, and the infiltration rate of surface soil was greater than those of the lower layers. ECs at the depths of 20–40 cm, 40–60 cm and 60–80 cm first decreased with time (before flowering stage) and then stabilized. From seeding stage to boll stage, the desalinization rates of the four soil layers were 91%, 61%, 40% and 34% from top to bottom of the profile, respectively. With an increase of irrigation times, the salt from the surface soil was leached downwards. Thus, subsurface pipe drainage was conducive to reducing soil salt and mitigating the adverse effect of soil salt on crop growth.

ECs in mildly salinized cotton field shows a different pattern of variation during growth stages and at different soil depths (Fig. 6). At the 0–20 cm depth, ECs during seeding stage was much higher than at other stages (P < 0.05). After seeding stage, ECs changed insignificantly (P > 0.05). At the 20–40 cm and 40–60 cm depths, ECs at seeding stage significantly differed from that of flowering stage (P < 0.05), and it was similar to the values of other stages (P > 0.05). Soil salinity at the 60–80 cm depth had no significant difference across various stages (P > 0.05). From seeding stage to flowering stage, soil salinity was effectively decreased at the depths of 0–20 cm, 20–40 cm and 40–60 cm, with the desalinization rates of 51%, 43% and 19%, respectively. After flowering stage, the phenomenon of salt gathering or salt returning occurred at the depths of 20–40 cm and 40–60 cm. At boll stage, only ECs at 0–20 cm decreased considerably compared to flowering stage (P < 0.05), and the desalinization rate of soil was 48%.

Soil salinity of the mildly and moderately salinized cotton field shows significant variation across different growth stages. The soil profile indicates the transition from surface accumulation to subsurface desalting. The depth of 60 cm from surface was a turnover point from significant desalinization to insignificant desalinization, even salt accumulation. Salinity of the surface soil in the moderately salinized plot was reduced to the level of mild salinization. For the mildly salinized plot, ECs of the surface soil decreased to the extent that salinization was alleviated. In relative term, the subsurface pipe drainage technique achieved a better improvement effect in moderately salinized plot.

Variation in drainage water EC_w and pH value

EC_w and pH values of drainage water from subsurface pipe are presented in Fig. 7. At the initial stage of irrigation (May 1–22), the irrigation amount was small, and the concentration of salts leached by irrigation was high. As the irrigation amount increased (between May 22 and July 10), the concentration of the leached salts gradually decreased due to dilution from the large volume of irrigation water. The concentration of salts leached from the farmlands increased because of decreasing irrigation amount between July 10 and September 6 (Fig. 4). After the last irrigation on September 6, EC_w remained stable due to lack of irrigation water. EC_w of the drainage water varied between 7.53 and 11.66 dS m-1 (i.e., −17%–27%). The pH value fluctuated within the range of 7.08–8.20, which was related to the irrigation amount, soil properties and drainage amount of subsurface pipes. The pH value in the whole drainage process was greater than 7, which indicates that water soluble salt ions (alkaline ions) were discharged and subsurface pipe drainage was conductive to the improvement of salinized soil. Soil salinity, soil texture, as well as the amount, frequency and quality of irrigation water, have certain influences on the desalinization effect. Further study needs to be undertaken for the determination of an optimal irrigation quota.

Subsurface pipe drainage can effectively reduce soil salinity and achieve the desalinization. Our findings are basically consistent with the previous studies (Christen & Skehan, 2001; Ritzema et al., 2008). The effect of the irrigation water was not explored, but would be expected to affect the desalinization of both the moderately and mildly salinized soils. The farmland irrigation water used in this study was from deep groundwater (depth: 300 m), which had a very low mineralization on average (EC: 0.35 dS m⁻¹). Therefore, the irrigation water had a little effect on salinity. During the irrigation stage (between May 1 and September 6), the average precipitation was 0.457 mm d⁻¹, while the average evapotranspiration was 3.659 mm d⁻¹. The low precipitation could not compensate for the high evapotranspiration, so the effect of precipitation on salt leaching was negligible. The reduction of salinity of the farmlands was mainly by leaching in irrigation water. Based on the samples taken at seeding stage before irrigation, on bud stage, flowering stage and boll stage, the results of soil profile analysis shows that salts were effectively leached by irrigation, and the salinity of all soil layers was decreased. The most substantial decrease occurred in the soil layer of 0–20 cm, from 6.26 dS m⁻¹ to 0.57 dS m⁻¹. Whole salinity at the depths of 40–60 cm and 60–80 cm decreased in bud stage and flowering stage. The average desalinization rate of the mildly salinized soil was 22%, and that of the moderately salinized soil was 56%. These results indicated that soil desalinization was closely related to subsurface pipe drainage.

Variation of cotton yield

The yield comparison between the two experimental plots with different levels of salinization (Fig. 8) indicates that there was no significant difference in cotton yield (P > 0.01). However, the differences were extremely significant when they were compared with the CK plots (P < 0.01). The yield on mildly salinized soil was 5,059 kg ha⁻¹, showing an increase of 25% compared with the corresponding CK plot. The yield increase was 55% in moderately salinized soil compared with the corresponding CK plot. Furthermore, stronger growth vigor was found in the treatments with subsurface pipe drainage than without subsurface pipe drainage, and the area of salinized farmlands decreased considerably. The emergence rate of cotton in the CK plots without subsurface pipe drainage was less than 20%, while that in the treatments with subsurface pipe drainage was 70% to 80% after first irrigation. Moreover, the death of seedlings after rain was reduced. These results indicate that crop yield increased significantly under the condition of subsurface pipe drainage.

Engineering economy

The cost of draining the open ditch is ¥ 45,000 for five years of normal operation with three pipes of controlled width equivalent to an open ditch (requiring bottom silt removal once a year); the pipes only require a one-off investment of ¥ 42,000 in the first year. Meanwhile, the concealed pipe does not occupy surface arable land, while the open ditch covers 15 acres per kilometer. Based on a benefit of ¥1,000 per acre, the piped drainage can save ¥ 75,000 over 5 years compared to open ditch drainage.