Water and Salt Balance in Agricultural Lands under Leaching with Shallow Subsurface Drainage Used in Combination with Cut-Drains

: Secondary salinization of irrigated lands in drylands is often caused by rising groundwater levels. Open drainage is widely employed to control groundwater. However, salinity levels tend to remain high under malfunctioning drainage conditions. Shallow subsurface drainage may be a possible solution to prevent salt accumulation, although it is di ﬃ cult for farmers to apply conventional tile drainage systems owing to construction costs. In this regard, we proposed a low-cost shallow subsurface drainage system used in combination with a new mole-drain drilling technology (cut-drain) developed in Japan, whose drainage capacity is similar to tile drain. The aim of this study is to evaluate the e ﬀ ect of the proposed system. The system was installed in a farmland, Uzbekistan. The experimental ﬁeld was set with / without the system to observe the di ﬀ erences in the balance of water and salt. The results revealed that the remaining inﬁltrated water in the ﬁeld decreased by approximately 26% and the removed net mass of salt was 14 Mg ha − 1 . The direction of salt movement changed from the deeper zone or surrounding ﬁeld to the open drainage. Therefore, the proposed system can enhance salt removal from ﬁelds.


Introduction
Since the middle of the 20th century, agricultural productivity in the drylands of Central Asia has increased with irrigation development. During this period, drainage systems have also been installed to sustain the productivity of irrigated lands [1,2]. However, secondary salinization of irrigated lands has been occurring through rising groundwater level (GWL) caused by inefficient irrigation and malfunctioning drainage systems [3][4][5]. Salinization, one of the most significant drivers of desertification, puts the irrigated agriculture at risk in the long-term [6,7]. In 2015, the United Nations proposed the Sustainable Development Goals (SDGs). It lists the sustainable agriculture through combating desertification and promoting restoration of the degraded soil as one (No. 15) of the goals [8]. Drainage technology has played an important role to realize the sustainable agriculture. In Syrdarya Province, the Syr-Darya, an international river flowing along the eastern edge of the Province, is a main water resource for irrigation and leaching. In this region, main and branch irrigation canals are concrete-lined. The depth of the main open unlined drainages, referred to as "collectors," is approximately 3.0-4.5 m in general, and that of the lateral open drainages is approximately 2.0 m or less. Furrow irrigation is mainly used in the region [23]. Before the irrigation or leaching period, temporal field ditches are constructed to distribute water in the field. The field gradient is almost zero, although some parts in the field are undulating.
This study was conducted on a farm (approximately 30 ha) under commercial cultivation located in the Axmedov Water Consumers' Association (WCA, a non-governmental organization for water users [24]), Mirzaobod District, an area that has the highest salinity level in the province. The experimental field in the farm is mapped in Figure 2. According to our topographical survey, the field gradient from east to west is approximately 1/900 on average. The soil is an alluvial soil and the textures are loam, sandy loam, and clay loam, according to the classification of the International Soil Science Society. The soil profile has a hard layer across 20-40 cm. The layer has a low hydraulic conductivity and a high bulk density (Table 1).

Layer
Ks Bulk Density cm s −1 g cm −1 Plow layer (0-20 cm) 1.3 × 10 −5 1.60 Hard layer (20- [25]. In the present study, due to the low permeability of the soil, the presence of community road/embankment along the drainages, the insufficient drainage depth, and the long horizontal distance between the drainage and the field, the hydraulic gradient from the field to the open drainage became too small to function effectively as the drainage system [26].

Experimental Set-Up
Cut-drain construction is mainly characterized by a unique drilling method. The drilling machine towed by a tractor is used to excavate a conduit to a certain depth (60-90 cm). A cavity made by the drilling machine acts as a relatively stable mole drain as shown in Figure 3. Cut-drains can act as supplemental subsurface drains when used as a connection to the main tile drain. The discharge capacity is no different to the cut-drain and typical tile drain. A cut-drain is constructed without any additional material. The soil texture most appropriate for cut-drain construction is one with a high clay and low sand content. In case of loamy soils, durability of it is short [27]. The design of the experimental field in this study having an area of 16 ha on the farm is illustrated in Figure 4. Later in the year, the field was used to produce cotton. Before sowing cotton, the experimental field was divided into 65 leaching plots (13 rows in an east-west direction with 5 plots in a north-south direction). The area with the SSDC was 4.0 ha. In the SSDC field, two main tile drains comprised of a perforated pipe and rice husk, were installed in 2015 at a depth of approximately 1.0 m. The diameter and length of the pipe were 10 cm and 200 m, respectively. Using a tractor of 130-140 HP, cut-drains were constructed in 2015 and 2016 to a depth of 60-90 cm with a spacing of 5 m. Since the cut-drain has a risk of preferential flow occurring which causes cavity collapse, we constructed it when the soil was wet along construction lines to mitigate the preferential flow in 2016 [28]. Cut-drains crossed the main Water 2020, 12, 3207 5 of 16 tile drains to connect parts of the rice husk. The length of the cut-drains was 100 m on the east side of the main tile drain. The direction of the cut-drains was parallel to the field gradient from east to west. The area of the control field was set near the SSDC field.      To observe soil moisture and salinity, soil samples in the cut-drain part and the control field were taken at the center of 15 plots and 9 plots, respectively, which were less affected by other factors such as tile drain or partial construction of cut-drain, and 4 points in the area near the drain pipe (a point 2.5 m distant from the pipe).
To observe fluctuation of groundwater depth, five observation wells using perforated pipes were installed to the depth of 4.5 m from the ground surface. Using these wells, the GWLs were monitored at each point in the control field (No. 1), the cut-drain part (No. 2), and the area near the drain pipe (No. 3). In addition, to compare the groundwater fluctuation between with and without leaching, two GWLs (No. 4 and No. 5) were also monitored. Point No. 4 was installed at a distance of 70 m from the experimental field in the same farm under winter wheat. No. 5 was installed at a distance of approximately 1.8 km from the experimental field, where there was no field under leaching nearby ( Figure 2). The GWLs were observed using water level loggers (Model HOBO-U20-001-04-TI, Onset Computer Corporation, Bourne, MA, USA) put into the wells. A barometer using the same model was also set near the field to calculate the GWL.

Water and Salt Balance
Kitamura et al. analyzed the water balance in an irrigated area of the lower Syr-Darya region of the Aral Sea Basin for a certain period [25]. We modified their water balance equation as follows: where ∆ is the difference in storage in the root zone to the depth of 100 cm between before and after leaching (mm), W is the amount of each process (mm). Subscripts s1 and s2 represent the storage in the soil before and after leaching, respectively, i is applied water for leaching, r is rain, g1 is groundwater inflow to the root zone, g2 is groundwater outflow from the outlet of the SSDC, g3 is seepage from the field to the open drainage, g4 is groundwater outflow from the root zone through paths other than g2 and g3, which is not emerged to open drainage but flowing vertically into deeper zone or horizontally into the neighboring dry field driven by hydraulic head, d is surface drainage water from the field, and et is the evapotranspiration (ET).
Since there is no surface drainage flow from the field, Wd can be disregarded and Equation (1) can be rearranged as follows: The left hand side is regarded as remaining water after leaching. By definition, Ws2 is the amount of water stored in the root zone (0-100 cm). Wg4 is outflow from this zone. Therefore, the remaining water does not include the initially stored water in the deeper zone or neighboring dry field.
Using Equation (2), the salt balance is expressed as follows: where S is the amount of salts (Mg ha −1 ), calculated by multiplying the amount of water (1 mm = 10 m 3 ha −1 ) and salt concentration (1 g L −1 = 10 −3 Mg m −3 ). Subscripts are the same as above. Note that the ET term is disregarded in the salt balance equation. The left hand side expresses a remaining salt mass after leaching. The effect of salt removal using the SSDC system is shown by comparison with the control field based on the calculation results of Equation (3).

•
Leaching water: Wi The average current velocity (m s −1 ) and its section area (m 2 ) were measured at upstream points in the temporal canals of the field once a day. The current velocity was measured using an electro-magnetic current meter (Model LP30, KENEK Corporation, Tokyo, Japan).

• Precipitation: Wr
Precipitation was measured at a farmer's garden located at a distance of 60 m to the north of the experimental field using a weather station (Model Vantage Pro2, Davis Instruments Corporation, Hayward, CA, USA).

•
Groundwater inflow to the root zone (0-100 cm) from outside: Wg1 Horizontal inflow toward the root zone can be disregarded because the field was surrounded by drainages and the surrounding field was relatively dry. Consequently, to determine Wg1, we presumed upward flow from groundwater toward the root zone.

•
Groundwater outflow from SSDC: Wg2 The discharge volume was estimated by measuring the volume poured from the outlet pipe using a bucket once a day.

•
Seepage from the field to the open drainage: Wg3 It was difficult to measure and determine the actual seepage from the experimental field to the open drainage. The seepage rate was estimated using a value which was derived by analyzing the composition of the open drainage water. In 2012, we observed the amount and salinity of open drainage, and the salinity of groundwater and irrigation water in Yangiobod WCA (Figure 1b) similar to Axmedov WCA, whose fields are not equipped with subsurface drainage systems, in Mirzaobod District. Simultaneously, we also obtained the amount of distributed irrigation water and precipitation from local organizations. Using these data, we estimated the seepage rate (Table S1) [29].
• Evapotranspiration: Wet Literature data for Syr-Darya River Basin, showing the reference evapotranspiration (ET0) every ten days [22], were applied to estimate the Wet.
• Salinity in the water: Si, Sr, Sg1, Sg2, Sg3 Samples of canal water used for leaching (Si), rain water (Sr), and the SSDC outflow (Sg2) were collected during the observed period. The sampling days of Si, Sr, and Sg2 were twice in the first leaching and once in the second, at four rainy days, and every day during the water supply and then approximately once every ten days in the later period, respectively. The ECw was measured using an EC meter (EC-33B, HORIBA Advanced Techno, Kyoto, Japan). Occasionally, total dissolved solids in the water (TDS, g L −1 ) was analyzed to estimate the mass of salts. Here, ECw of the other groundwater (Sg1 and Sg3) is also assumed to be the same as that of the SSDC outlet water.
• Salinity in the soil: Ss1, Ss2 The electrical conductivities of a 1:1 suspension (EC1:1) of sampled soils were measured. Each EC1:1 value was converted to ECe using an empirical formula presented by the Research Institute of Irrigation and Water Problems of Uzbekistan [30]. To estimate the salt balance, TDS volume in the soil (TDSsoil), expressed in terms of mass percentage, before and after leaching was analyzed. TDSsoil is determined by a percentage of TDS weight in 1:5 suspension to the dry soil weight. The change in soil salinity was evaluated with an analysis of variance method.

•
Soil moisture: Ws1, Ws2 Soil samples were taken before leaching on 19 and 20 December 2016, and after leaching on 27 and 28 February 2017. The volumetric water contents before and after leaching are shown in Figure 5. Soil moisture was horizontally uniform before leaching while those of surface and bottom layers were higher than that of the middle layer. During leaching, all plots were flooded, and at the end of February    The stored water in the root zone before and after leaching was 261 and 384 mm in the control field, and 259 and 379 mm in the SSDC field, respectively.

•
Leaching water: Wi Leaching in the experimental field was carried out twice according to the WCA's water distribution plan. The first leaching was conducted from 21 at 11 a.m. to 29 at 11 a.m. December 2016, and the second from 3 at 11 a.m. to 12 at 4 p.m. January 2017. In the SSDC field, leaching was carried out from 25 at 10 a.m. to 27 at 9 p.m. December 2016, and then restarted from 6 at 6 p.m. to 11 at 6 p.m. January 2017. The depths of the first and second leaching were 192 and 118 mm, respectively, resulting in a 310 mm of total leaching.

•
Precipitation: Wr The precipitation was 127 mm during the 70 days.
• Groundwater inflow to the root zone from outside: Wg1 The changes in GWL are shown in Figure 6. Before leaching, all GWLs had similar rising rates. The GWLs at No. 1-3 suddenly rose up to the depth of 10-60 cm during the first leaching, then, repeated up and down in the zone of 0 to 100 cm. The changes in GWL were influenced by leaching and rain events. The GWL at No. 4 slightly enhanced the rising after leaching until the middle of January, then, rising rate increased, and finally synchronized to the other wells in the field at the beginning of February. The GWL of No. 4 seems to be linked to the leaching. The GWL at No. 5, where there was no field under leaching nearby, kept an equal rising rate at approximately 10 mm per day until the end of January. This rising rate might be a natural fluctuation. Excluding rains, the rising rate was nearly constant until the end of February. Applying this rising rate of No. 5 to No. 1-4, the GWLs would not have reached to the depth of 100 cm during the 70 days if leaching had not been carried out. Consequently, we may neglect the upward groundwater inflow into the root zone (i.e., Wg1 = 0).   • Groundwater outflow from SSDC: Wg2 The discharge rate at the outlet is shown in Figure 7, along with leaching water and precipitation. The outlet was closed during the first leaching period to ensure sufficient time to soak the soil with leaching water. Nine hours after completion of water supply in the first leaching, discharge started. The outlet was opened on 30 December after a lapse of sixty hours from the completion of the leaching, and measurement of the discharge rate was commenced. Therefore, discharge rate measured on the first day also included the discharge accumulated from previous two days. Although some data were missing, the total discharge for the 70 days was 149 mm, corresponding to 34% of the applied water. Figure 7. Amounts of discharged water and input water in the SSDC field. The horizontal axis is date and the upper side shows the input water for leaching and precipitation and the lower shows the discharged water from the SSDC outlet. The asterisks indicate missing data in the discharged water.

•
Seepage from the field to the open drainage: Wg3 Because the input water during the period was 437 mm, the amount of groundwater outflow by seepage (Wg3) was estimated to be 52 mm, assuming that the same portion, 12%, of applied water seeped out. It was approximately one-third of the groundwater outflow from the SSDC (Wg2). The lower seepage amount may be attributed to the low hydraulic gradient. •

Evapotranspiration: Wet
According to Bezborodov et al. [22], the ET0 ranges between 0.4 and 0.8 mm d −1 in December and January, and 1.0 and 1.5 mm d −1 in February in Syr-Darya River Basin. Assuming that the evaporation rate is the same as ET0 during leaching period in the rainy season, the total ET for the 70 days was approximately 56 mm.
• Salinity in the water: Si, Sr, Sg1, Sg2, Sg3 The average ECw of leaching water for Si, rain water for Sr, and the discharged water from the SSDC for Sg2 during the 70 days were 1.2, 0.3, and 10.2 dS m −1 , respectively. The ECw was converted to the total salt mass in the water using an empirical formula determined using the data collected on the farm (Figure 8a,b). The total salt of input (Si and Sr) and discharged water (Sg2) was 3.3 and 17.2 Mg ha −1 during the 70 days, respectively, indicating that 13.9 Mg ha −1 of salts could be removed through the SSDC. The upward salt flux to the root zone could be neglected due to the absence of upward flow to the root zone (i.e., Sg1 = 0). Regarding the seepage, we assumed that the seeped water had the same ECw as the discharged water from the outlet of the SSDC, resulting in 6.0 Mg ha −1 of Sg3.   • Salinity in the soil: Ss1, Ss2 The average ECe analyzed from the upper (0-60 cm) and lower layer (60-100 cm) before and after leaching are shown in Figure 9. The figures above the bars show the ECe and the figures in parentheses show the percentage of the ECe after leaching compared with the corresponding value before leaching. The relative changes in soil salinity in the control field, the cut-drain part, and the area near the drain pipe were 67%, 62%, and 50% in the upper layer, and 85%, 87%, and 62% in the lower layer, respectively. There were no significant differences under the analysis of variance. Obtained ECe values were converted to TDSsoil using an empirical formula shown in Figure 8c. We found that the total salt mass in the root zone changed before and after leaching were 93.4 and 69.7 Mg ha −1 in the control field, and 85.4 and 60.4 Mg ha −1 in the SSDC field, respectively. Therefore, the total reduction in salt mass owing to leaching in the control and the SSDC field were 23.7 and 25.0 Mg ha −1 , respectively.

Water and Salt Balance
• Remaining water in the fields: Ws2 + Wg4 The result of water balance calculation is listed in Table 2. The remaining amount of water, which was not discharged to open drainage, was 590 mm for the control field and 439 mm for the SSDC field. The main difference in remaining water occurred in the Wg4. The Wg4 of the SSDC field was 71% lower than that of the control field. Note: Observation period is from 20 December 2016 to 27 February 2017, for 70 days. Abbreviations: ET is evapotranspiration. W and S represent water amount (mm) and salt mass (Mg ha −1 ), respectively. Subscripts i is leaching water, r is rain, g1 is groundwater inflow to the root zone, g2 is groundwater outflow from the outlet of the SSDC, g3 is seepage from the field to the open drainage, g4 is groundwater outflow from the root zone through paths other than g2 and g3, et is the evapotranspiration, s1 and s2 represent the storage in the soil before and after leaching. The salt mass of Ss1 and Ss2 is the sum of the salt mass contained within the cubic volume in each 20 cm of depth, which is derived by multiplying the TDSsoil (%) with the bulk density (g cm −3 ) of each root zone.
According to the GWL observation, the leaching raised the GWL. Furthermore, the observation well No.4 showed that the infiltrated water reached the drier neighboring field. The lower Wg4 means a reduction in the downward percolation. Thus, the SSDC may have restricted groundwater recharge in wider parts of the field.

•
Remaining salt in the fields: Ss2 + Sg4 The result of salt balance calculation is also listed in Table 2. The remaining salt, not disposed of to open drainage, was 90.7 and 65.5 Mg ha −1 for the control field and the SSDC field, respectively. The Ss1 in the root zone of the SSDC field was 8.0 Mg ha −1 less than that of the control field. The difference widened after leaching. The SSDC reduced salt mass a little in the root zone, while the difference in the Sg4 reached 15.9 Mg ha −1 , 76% lower than that of the control field. In the same way as water, the lower Sg4 for the SSDC field implies that the SSDC restricts the contamination of surrounding fields with salts through horizontal outward flow from the field under leaching.

Water and Salt Balance
Calculated water and salt balance indicate that the SSDC system could enhance the discharge of infiltrated water and salt. The discharged water was highly saline in nature. The differences in ECw between leaching and discharged water indicate that the SSDC removed accumulated salts, showing that the SSDC could assure discharge of groundwater containing much salt even under limited drainage conditions. Discharge from the SSDC was also observed after rain events in March and April when the GWL was high. Therefore, the SSDC system could also promote discharge process under natural precipitation.
Due to enhanced outflow by the SSDC, the remaining water in the SSDC field decreased by 26% compared with that of the control field. The major portion of this decrease may be due to percolation from the root zone to the deeper zone or neighboring dry field. In other words, the SSDC caught the percolating water and discharged it to the open drainage. Likewise, the remaining salt mass was analogous to the remaining water decrease. The soil salinity had large variation in the control field, the cut-drain part and the area near the drain pipe, respectively, making the difference statistically insignificant. As shown in Figure 9, the upper layer in the cut-drain part and the area near the drain pipe lost more salts than that in the control field. In the root zone, the differences of salt mass between the control field and the SSDC field were small. The leaching would have pushed much salt out of the root zone in both fields. Even under the circumstance of high GWL with high salinity, the famer has practiced leaching to push the salt out of the surface layer, however, the excluded salt remains in the deeper zone or surrounding field due to the limited seepage of groundwater to the open drainage. After leaching, the remaining salt steadily rises to the surface layer with upward flow of water. The reduction in remaining water and salt below the root zone could reduce re-accumulation of salt. The SSDC system was found to fulfill the role for the reduction.

Future Development of the SSDC
The SDGs urges combating desertification and halting and reversing land degradation for sustainable agriculture. To combat desertification, various transitions in the current soil and water system are needed [31]. Innovations in drainage technology are also required to realize the SDGs by 2030. Subsurface drainage is one of the measures to overcome a serious and increasing salinization issue for sustainable development of irrigated agriculture in drylands [32]. It needs more studies for cost reduction and technology sophistication. Viewed in this light, the cut-drain can be an alternative for famers in developing countries because of easy construction using a tractor without any material. The issue of short durability caused by the characteristic of not using reinforcing materials could be solved through incorporating cut-drains into the farming system as a preparation work of the leaching. In addition to that, as a result of the yield survey at the experimental field in September 2017, the cotton yield in the control field, cut-drain part, and near the drain pipe were 3.3, 3.9, and 4.0 Mg ha −1 , respectively ( Figure S1). The SSDC system gave approximately 18% higher yield in average compared with the control [26].
To make the effect of the SSDC system more reliable, there are some issues that remain to be resolved. Further studies are needed to clarify the long-term effects, as well as negative effects relating to GWL. Secondly, in this study, seepage rate from the field to the open drainage were estimated based on the observed seepage rate by previous field data in a similar WCA. Thus, there might be uncertainty in actual seepage rates including the difference of seepage volumes between the control field and the SSDC field. To analyze water and salt balance more accurately, it would also be necessary to accumulate more drainage discharge data from various fields. Finally, further field experiments with plural repetitions to enable statistical analysis may be required in near future.

Conclusions
Salinization is one of the drivers of desertification. It is often caused by rising GWL. Drainage systems have been installed to control GWL and leaching is practiced to remove salts from the field. However, the drainage systems are in a state of disrepair in some areas where salinity levels remain high under malfunctioning drainage conditions. To remove accumulated salt, we suggest a low-cost shallow subsurface drainage system used in combination with cut-drains (SSDC). It could be an alternative subsurface drainage technology in drylands. The aim of this study was to evaluate the effect of the system. The experimental field was set with/without the system to observe the differences in the water and salt balance on a farmland in Syrdarya Province, Uzbekistan. It is the first trial to evaluate the effect of cut-drains on shallow subsurface drainage as a salt removal technology in drylands.
The results reveal that the SSDC could reduce accumulated salts, removing 14 Mg ha −1 of salts in net mass, down to approximately between one-third and one-fourth of remaining salt of the control field. The reduction in salts in the root zone was slightly larger than the control field. The most beneficial part of the reduction could be in the deeper zone or surrounding fields. The reduction in remaining salt under the SSDC system could reduce re-accumulation of salt. This implies that the technology can be a means to enhance salt removal from the field. The low-cost and easy construction of the cut-drain can be an advantage for farmers. Although more studies are necessary in terms of long-term effect and more accurate water and salt balance analysis, we expect that the SSDC will contribute to mitigating salinization and benefiting farmers who struggle with the difficulty of salinization. Funding: This study was partially funded by an overseas agricultural research project under the Ministry of Agriculture, Forestry and Fisheries, Japan. The study work was conducted as a research project in the Resource Management in Asia and Pacific Islands by the Japan International Research Center for Agricultural Sciences.
Acknowledgments: We gratefully acknowledge the Farmers' Council of Uzbekistan as a research counterpart, the Research Institute of Irrigation and Water Problems of Uzbekistan, the Hydro-Geology Melioration Expedition and farmers in Syrdarya Province, Uzbekistan. We would also appreciate all the parties concerned for their support to this study.

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript.

DWL
Drainage water level ECe Electric conductivity of a saturated extract EC1:1 Electric conductivity of a 1:1 suspension ET Evapotranspiration ET0 Reference evapotranspiration GWL Groundwater level SDGs Sustainable Development Goals

Sg1
Salt mass in groundwater inflow to the root zone (0-100 cm) Sg2 Salt mass in groundwater outflow from the outlet of the SSDC Sg3 Salt mass in groundwater outflow by seepage from the field to the open drainage

Sg4
Salt mass in groundwater outflow from the root zone (0-100 cm) into the deeper zone or neighboring field Si Salt mass in irrigation water Sr Salt mass in rain water Ss1 Salt mass in root zone (0-100 cm) before leaching Ss2 Salt mass in root zone (0-100 cm) after leaching SSDC Shallow subsurface drainage with cut-drains TDS Total dissolved solid in the water TDSsoil Total dissolved solid in the soil WCA Water Consumers' Association Wd Amount of water of surface discharge water from the field Wet Amount of water of evapotranspiration Wg1 Amount of groundwater inflow to the root zone (0-100 cm) Wg2 Amount of groundwater outflow from the outlet of the SSDC Wg3 Amount of groundwater outflow by seepage from the field to the open drainage

Wg4
Amount of groundwater outflow from the root zone (0-100 cm) into the deeper zone or neighboring field Wi Amount of irrigation water Wr Amount of rain water Ws1 Amount of water in root zone (0-100 cm) before leaching Ws2 Amount of water in root zone (0-100 cm) after leaching