Watercourse Pollution due to Surface Runoff following Slurry Spreading. Part I: Calibration of the Soil Water Simulation Model SOIL for Fields Prone to Surface Runoff

The Swedish soil water model SOIL has been calibrated for several drained fields in Scotland and Ireland. Drainage efficiency in these fields varies, with inefficient drainage systems leading to saturated profiles and large surface runoff flows. The model has been modified to represent drainflow in typical Scottish and Irish fields in which permeable backfill extending to the surface is present directly above plot drains. When the conductivity of the backfill material is low, surface runoff is shown to be enhanced in specific soil types. Overall, the predictions of the modified model are in reasonably good agreement (as shown by the efficiency factor values) with measured water table levels, drain and surface runoff flows in these fields. These calibrated fields are to be used in subsequent work on pollution from surface runoff following slurry spreading. A useful indicator of potential runoff risk in such systems is the total saturated hydraulic conductivity of the profile, defined here as arising from the combination of the saturated soil and drain conductivities. Fields are classified into high risk if the conductivity of the profile is lower than 6 mm/d, low risk if the conductivity is greater than 18 mm/d, and moderate risk for intermediate conductivities. A sensitivity analysis of the model with regard to drain and surface runoff flows, varying the drain spacing, a backfill resistance term, the soil matrix and macropore saturated hydraulic conductivities, soil porosity and the pore size distribution index, is also presented. This analysis shows that in order of increasing importance, backfill resistance, macropore saturated hydraulic conductivity and drain spacing, have the largest effect on the generation of surface runoff.     

A model for the design of drainage in flat agricultural lands

A model is presented that can be used to determine drainage measures and their costs. It has been elaborated for a wet tropical climate, for situations with open field drains, shallow groundwater table and a homogenous soil underlain by an impervious layer. The land is flat and the proposed agricultural use requires control of the groundwater table.A basic element of the model is a scheme to compute the water balance per day for a drainage parcel. Discharge, evapotranspiration, groundwater level and soil moisture storage are estimated as functions of rainfall, potential evapotranspiration, vegetation and soil characteristics and of an assumed drainage intensity. The water balance computation is performed for periods of 5–40 years of daily rainfall data, for a series of drainage intensities. The results can be subjected to a drainage criterion, from which a design drainage intensity and a corresponding drain spacing can be derived.Finally the layout of canals for a block of 4 × 1 km² is determined and excavation and a series of canal characteristics are computed.A summary of some applications is included.     

Time-dependent drainage from root zone and drainage coefficient under different irrigation management levels for subsurface drainage design in Egypt

Drainage water from the lower boundary of the root zone is an important factor in the irrigated agricultural lands for prediction of the water table behavior and understanding and modeling of water and chemical movement in the soil profile. The drainage coefficient is an important parameter for the design of subsurface drainage. On a 33,138 ha of the Nile Delta in Egypt, this study is conducted using 90 irrigation periods over a 3-year crop rotation to estimate the time-dependent drainage from the root zone and the design subsurface drainage coefficient with different cropping seasons and irrigation management levels.The results showed that the cropping seasons and the irrigation management levels as indicated by different irrigation efficiency are significantly affected the drainage rate from the root zone and the design value of subsurface drainage coefficient. Drainage rates from the root zone of 1.72 mm/d and 0.82 mm/d were estimated for summer and winter seasons, respectively. These rates significantly decreased in a range of 46% to 92% during summer season and 60% to 98% during winter season when the irrigation efficiency is increased in a range of 5% to 15%. The subsurface drainage coefficient was estimated to be 1.09 mm/d whereas the design drain pipe capacity was estimated to be 2.2 mm/d, based on the peak discharge of the most critical crop (maize), rather than 4.0 mm/d which is currently used. A significant decrease of the drainage coefficient and the drain pipe capacity ranges from 18% to 45% was found with the increase of irrigation efficiency in a range of 5% to 15%. The leaching requirement for each crop was also estimated.     

Drainage and desalinization of heavy clay soil in Portugal

The Leziria Grande area consists mainly of poorly drained, saline clay soils of marine origin. Three experimental fields were laid out to find whether subsurface drainage can be effective in lowering the groundwater table and improving desalinization.Subsurface drainage results in a lower groundwater table than does surface drainage. With increasing spacing, the groundwater remains at a higher level for longer periods, which is expressed here by the sum of exceedances of the groundwater table above 30 cm during winter.Soil salinity, expressed as EC_1:2, and sodicity, expressed as E.S.P., decreased during the first 3 years, in which precipitation varied between 600 and 750 mm and the average drain outflow was about 250 mm. The leaching efficiency decreased with time, indicating that the removal of salt is a slow process in fine-textured soil.Application of gypsum lowered the E.S.P. The infiltration rate and the drain outflow increased. Although the total amount of salts in the drainwater was 40% higher than for the untreated plots, no lower EC_1:2 values were found. This is ascribed to spatial variability in soil salinity.     

Using the DRAINMOD-N model to study effects of drainage system design and management on crop productivity,profitability and NO3–N losses in drainage water

The environmental impacts of agricultural drainage have become a critical issue. There is a need to design and manage drainage and related water table control systems to satisfy both crop production and water quality objectives. The model DRAINMOD-N was used to study long-term effects of drainage system design and management on crop production, profitability, and nitrogen losses in two poorly drained soils typical of eastern North Carolina (NC), USA. Simulations were conducted for a 20-yr period (1971–1990) of continuous corn production at Plymouth, NC. The design scenarios evaluated consisted of three drain depths (0.75, 1.0, and 1.25 m), ten drain spacings (10, 15, 20, 25, 30, 40, 50, 60, 80, and 100 m), and two surface conditions (0.5 and 2.5 cm depressional storage). The management treatments included conventional drainage, controlled drainage during the summer season and controlled drainage during both the summer and winter seasons. Maximum profits for both soils were predicted for a 1.25 m drain depth and poor surface drainage (2.5 cm depressional storage). The optimum spacings were 40 and 20 m for the Portsmouth and Tomotley soils, respectively. These systems however would not be optimum from the water quality perspective. If the water quality objective is of equal importance to the productivity objective, the drainage systems need to be designed and managed to reduce NO₃–N losses while still providing an acceptable profit from the crop. Simulated results showed NO₃–N losses can be substantially reduced by decreasing drain depth, improving surface drainage, and using controlled drainage. Within this context, NO₃–N losses can be reduced by providing only the minimum subsurface drainage intensity required for production, by designing drainage systems to fit soil properties, and by using controlled drainage during periods when maximum drainage is not needed for production. The simulation results have demonstrated the applicability of DRAINMOD-N for quantifying effects of drainage design and management combinations on profits from agricultural crops and on losses of NO₃–N to the environment for specific crop, soil and climatic conditions. Thus, the model can be used to guide design and management decisions for satisfying both productivity and environmental objectives and assessing the costs and benefits of alternative choices to each set of objectives.     

Subsurface drainage to combat waterlogging and salinity in irrigated lands in India: Lessons learned in farmers’ fields

The introduction of irrigated agriculture in the arid and semi-arid regions of India has resulted in the development of the twin problem of waterlogging and soil salinization. It is estimated that nearly 8.4 million ha is affected by soil salinity and alkalinity, of which about 5.5 million ha is also waterlogged. Subsurface drainage is an effective tool to combat this twin problem of waterlogging and salinity and thus to protect capital investment in irrigated agriculture and increase its sustainability. In India, however, subsurface drainage has not been implemented on a large scale, in spite of numerous research activities that proved its potential. To develop strategies to implement subsurface drainage, applied research studies were set-up in five different agro-climatic sub-regions of India. Subsurface drainage systems, consisting of open and pipe drains with drain spacing varying between 45 and 150 m and drain depth between 0.90 and 1.20 m, were installed in farmers’ fields. The agro-climatic and soil conditions determine the most appropriate combination of drain depth and spacing, but the drain depths are considerably shallower than the 1.75 m traditionally recommended for the prevailing conditions in India. Crop yields in the drained fields increased significantly, e.g. rice with 69%, cotton with 64%, sugarcane with 54% and wheat with 136%. These increases were obtained because water table and soil salinity levels were, respectively, 25% and 50% lower than in the non-drained fields. An economic analysis shows that the subsurface drainage systems are highly cost-effective: cost-benefit ratios range from 1.2 to 3.2, internal rates of return from 20 to 58%, and the pay-back periods from 3 to 9 years. Despite these positive results, major challenges remain to introduce subsurface drainage at a larger scale. First of all, farmers, although they clearly see the benefits of drainage, are too poor to pay the full cost of drainage. Next, water users’ organisations, not only for drainage but also for irrigation, are not well established. Subsurface drainage in irrigated areas is a collective activity, thus appropriate institutional arrangements for farmers’ participation and organisation are needed. Thus, to assure that drainage gets the attention it deserves, policies have to be reformulated.     

石羊河流域井灌区土壤水分深层渗漏研究

基于土壤水分亏缺、总有效水分和实际有效水分的概念,建立土壤水分水量平衡模型,通过计算作物根系层的深层渗漏量,来反映土壤水与地下水之间量的相互转化关系。研究结果表明,研究期间的总深层渗漏量为9 4.1 9 mm,占研究期间总灌水量和降雨量的2 0.1%,这部分水量通过根系层补给下层土壤,最终补给地下水。     

Rainfall probability forecasts used to manage a subdrainage-subirrigation system for watertable control

Water management of a subsurface-drainage and subirrigation system was simulated using a daily rainfall probability index (rpi), to control the watertable depth (wt) in the soil profile. Daily management of ‘free drainage’, controlled drainage, or subirrigation, was based upon the rpi value. The rpi was computed from the daily rainfall probability in forecasts issued by the U.S. National Weather Service. Climatic data and weather forecast records (1979–1985) for the lower Mississippi Valley were used in the DRAINMOD program to simulate daily fluctuations in the watertable. Various statistical and summation equations were used for computing the rpi. Management success was evaluated by conditions of excess and deficit soil water in the root zone, and by predicted crop yield. Using only the ‘today’ and ‘tonight’ segments of the morning (5:25 h) forecast, 75% of the significant rainfall events occurring during the growing season were successfully predicted when the rpi ≥ 0.60. Free drainage in advance of predicted storms significantly reduced the duration of excess soil water in the root zone and increased simulated maize yield 0 to 11%, when compared to controlled drainage where the water level at the drain outlet was maintained constant at a level above the drain.     

农田不同排水措施地下排水效果模拟试验研究

在地下水位较高、地表易于形成积水的中国南方地区,通过农田排水措施可以及时排除多余地表积水,快速降低地下水位,以达到排涝降渍、协同调控的目的.文中基于室内砂槽试验,揭示暗管排水、明沟排水、不同反滤体高度的反滤体排水及改进暗管排水等措施的地下排水规律及效果.结果表明:将暗管周围土体置换为高渗透性土体介质的改进暗管排水可明显提高排水流量,当土体置换高度达2 cm时(对应于田间条件40 cm),其排水流量均高于相同埋深条件下的其他排水措施,达暗管排水的1.59~1.66倍;改进暗排在地表积水消失时仍保持较大的排水流量,可达相同埋深暗管流量的2倍以上,在积水层消失后,能迅速降低农田土壤水的渍害胁迫,将地下水位降低至暗管埋设高度;各种排水措施,在地表积水即将消失时,出现了流量与水头变化幅度较大的现象.相对于各种地下排水措施,改进暗管排水在除涝降渍中存在明显优势.研究结果可为涝渍灾害易发地区高效除涝降渍减灾工程设计和建设提供参考.     

Predicting effects of drainage water management in Iowa's subsurface drained landscapes

Long-term hydrologic simulations are presented predicting the effects of drainage water management on subsurface drainage, surface runoff and crop production in Iowa's subsurface drained landscapes. The deterministic hydrologic model, DRAINMOD was used to simulate Webster (fine-loamy, mixed, superactive, mesic) soil in a Continuous Corn rotation (WEBS_CC) with different drain depths from 0.75 to 1.20 m and drain spacing from 10 to 50 m in a combination of free and controlled drainage over a weather record of 60 (1945-2004) years. Shallow drainage is defined as drains installed at a drain depth of 0.75 m, and controlled drainage with a drain depth of 1.20 m restricts flow at the drain outlet to maintain a water table at 0.60 m below surface level during the winter (November-March) and summer (June-August) months. These drainage design and management modifications were evaluated against conventional drainage system installed at a drain depth of 1.20 m with free drainage at the drain outlet. The simulation results indicate the potential of a tradeoff between subsurface drainage and surface runoff as a pathway to remove excess water from the system. While a reduction of subsurface drainage may occur through the use of shallow and controlled drainage, these practices may increase surface runoff in Iowa's subsurface drained landscapes. The simulations also indicate that shallow and controlled drainage might increase the excess water stress on crop production, and thereby result in slightly lower relative yields. Field experiments are needed to examine the pathways of water movement, total water balance, and crop production under shallow and controlled drainage in Iowa's subsurface drained landscapes.     

Evaluating the Impact of Irrigation and Drainage Policies on Agricultural Sustainability

Farmers in the Broadview Water District in central Californiahave been improving irrigation practices in response to risingirrigation water prices and reductions in water supply since1989, when incentive policies were first implemented to reducethe volume of subsurface drain water generated in theDistrict. The average salinity of water deliveries hasincreased, over time, as the District has recycled largeamounts of drainage water to achieve regional restrictions ondrainage water discharge. We review irrigation and drainageactivities in Broadview since 1986 with an emphasis on thesustainability of crop production when drainage discharge islimited. Average cotton yields in Broadview have declined inrecent years, both nominally and in comparison with averageyields reported for the large county in which Broadview islocated. Average tomato yields in Broadview have increased inrecent years, but county-wide yields have exceeded Broadviewyields with greater frequency than in the late 1980s. Theseobservations suggest that average crop yields in Broadview maybe starting to reflect the increasing salinity of soil andwater resources, which may be due in part to persistentrestrictions on drainage water discharge.     

Changes in spatial and temporal variability of SAR affected by shallow groundwater management of an irrigated field, California

In the irrigated western U.S. disposal of drainage water has become a significant economic and environmental liability. Development of irrigation water management practices that reduce drainage water volumes is essential. One strategy combines restricted drainage outflow (by plugging the drains) with deficit irrigation to maximize shallow groundwater consumption by crops, thus reducing drainage that needs disposal. This approach is not without potential pitfalls; upward movement of groundwater in response to crop water uptake may increase salt and sodium concentrations in the root zone. The purposes for this study were: to observe changes in the spatial and temporal distributions of SAR (sodium adsorption ratio) and salt in a field managed to minimize drainage discharge; to determine if in situ drainage reduction strategy affects SAR distribution in the soil profile; and to identify soil or management factors that can help explain field wide variability. We measured SAR, soil salinity (EC_1:1) and soil texture over 3 years in a 60-ha irrigated field on the west side of the San Joaquin Valley, California. At the time we started our measurements, the field was beginning to be managed according to a shallow groundwater/drainage reduction strategy. Soil salinity and SAR were found to be highly correlated in the field. The observed spatial and temporal variability in SAR was largely a product of soil textural variations within the field and their associated variations in apparent leaching fraction. During the 3-year study period, the percentage of the field in which the lower profile (90-180 cm) depth averaged SAR was above 10, increased from 20 to 40%. Since salinity was increasing concomitantly with SAR, and because the soil contained gypsum, sodium hazard was not expected to become a limiting factor for long term shallow groundwater management by drain control. It is anticipated that the technology will be viable for future seasons.     

Modelling the water balance of effluent-irrigated trees

Irrigation of effluent is an increasingly popular treatment option due to concern about nutrient additions to rivers and coastal waters. Since some studies have shown that irrigation with waste water can lead to contamination of groundwater resources, there is need for a model to predict the fate of irrigated water, salt, and nitrogen that can be applied to a variety of different soils, climates, and crops. We present the development of the water balance part of such a model, APSIM for Effluent, and carry out a comparison against data obtained from an effluent-irrigated plantation of Eucalyptus grandis. Over 10 months, modelled tree water use was within 1.5% of that obtained by sap-flux measurements. When compared over 5 years of the experiment, modelled drainage lay above that estimated by a water balance technique, which was known a priori to underestimate drainage, and was close to that estimated by the chloride mass balance technique. Simulated chloride accumulated in the soil was within the scatter of the observations, although it was consistently at the lower end of the range of the data. There was good agreement between the model predictions and measured chloride concentration distribution with depth in the soil. A considerable amount of water was lost as deep drainage, even for the treatment that aimed to add only enough effluent to replace that lost by evaporation. During 5 years, of the 3370 mm rainfall and 4480 mm effluent received by that treatment, 6710 mm was lost by the various evaporative routes, and 1080 mm was lost by deep drainage.     

Irrigation application efficiency and deep drainage potential under surface irrigated cotton

Furrow irrigation events conducted under usual farmer management were analysed to determine the irrigation application efficiencies being attained, and the magnitude of the irrigation contribution to deep drainage under surface irrigated cotton in Queensland. Application efficiencies were shown to vary widely from 17 to 100% and on average were a low 48%. Losses to deep drainage were substantial, averaging 42.5 mm per irrigation. This has the potential for significant environmental harm and also represents an annual loss of up to 2500 m³/ha (2.5 Ml/ha) of water that could be beneficially used to grow more cotton. Simulations of each event using the simulation model SIRMOD illustrated simple ‘recipe’ strategies that would lead to gains in efficiency and reductions in the deep drainage losses. Additional simulations of selected events showed that further significant improvements in performance can be achieved by the application of more advanced irrigation management practices, involving in-field evaluation and optimisation of the flow rate and irrigation time to suit the individual soil conditions and furrow characteristics. Application efficiencies in the range 85–95% are achievable in all but the most adverse conditions. The dependency between deep drainage and irrigation management was demonstrated, confirming that substantial reductions in deep drainage are possible by ensuring that irrigation applications do not exceed the soil moisture deficit.     

Water use,water-use efficiency and yield of dryland chickpea as influenced by P fertilization,stored soil water and crop season rainfall

Dryland chickpea is grown on stored water in the soil profile and with limited crop season rainfall (CSR). In a field experiment, carried out for 3 years on silty loam soil, water extraction pattern, water use and its efficiency by chickpea in relation to P application, stored soil water and crop season rainfall have been investigated. Stored soil water varied from 182 to 246 mm in a meter profile and the CSR varied from 72 to 184 mm.Response of P application increased with increasing initial water storage in the soil profile. The first 60–100 days of crop growth appeared to be the most critical. Water stress during this period severely affected the yield. Rainfall after 100 days did not appear to have been fully utilized by the crop, especially when the crop had already suffered from water stress between 60 and 100 days. Compared with the control, P application increased yield, water use and water-use efficiency. Soil water depletion was 25% greater for the fertilized crop than for the unfertilized crop.     

Evaluation of DRAINMOD for predicting water table heights in irrigated fields at the Jordan Valley

A detailed field experiment was carried out in the Jordan Valley, south of Lake Kinneret, Israel for evaluation of the water management model DRAINMOD. This field was chosen to represent the local agro-climate conditions of that zone. Banana crop was grown and was irrigated daily with about 3200 mm/year and 0.5 leaching fraction. Subsurface drainage system with 2.5 m drain depth and 160 m drain spacing existed in the field. The water table depth was measured with about 100 piezometers, in which most of them were observed weekly, and four were continuosly recording piezometers. Five identical drainage plots were selected, out of 10 existing, as replicates for the evaluation of DRAINMOD. Deviations in a range of 0.3–1.7 m between observed water table depth and that simulated by DRAINMOD were found in four out of the five replicates. A reasonable agreement was found only in one drainage plot out of the five tested. These findings contradict the world wide convention that DRAINMOD simulation is in a good agreement with observed field data. An additional study was therefore conducted to explore the reasons for these large deviations. Three reasons were suggested: (i) a strong side effect by the Jordan River, which flows some 350 m west to the test field; a very steep 4.6% gradient was found toward the Jordan River; (ii) presence of sandy permeable layers below the depth of the drains which magnifies the boundary condition effect of the Jordan River; (iii) a very significant component of deep and lateral seepage (more than 50% of the yearly irrigation plus rainfall). A combination of these three reasons was suggested as an explanation to the apparent large disagreement. It was therefore recommended not to use DRAINMOD or similar vertical flow models for simulation of water table depths in irrigated fields with subsurface drain pipe systems in the Jordan Valley.     

滴灌条件下暗管滤层结构对排水、排盐效果的影响

为解决滴灌农田非饱和条件下暗管排水困难的问题,设置2种暗管滤层铺设方式,以常规暗管滤层进行排水为对照,基于室内土槽试验,分析了滤层铺设方式对暗管排水排盐效果的影响及其机理。T1处理为常规暗管滤层铺设方式,暗管四周铺设细砂滤层,T2处理为暗管上部铺设细砂,T3处理为细砂斜垫层斜铺连接体积质量分界层与暗管。结果表明,T1处理受土壤水滞后效应影响显著,暗管不排水,土壤水、盐积聚于暗管底部;T2、T3处理可在暗管上部产生局部饱和区,促使暗管排水。T2处理排水时所需历时较长,排出的水盐总量较少;T3处理可使暗管最早排水,排水时暗管下部土壤积盐最少,排水流量和排水盐总量最大。     

农田排水沟水土流失成因与防治

黑龙江省本着改造中低产田,夺取农业高产的原则,对易受涝灾和渍害的田地进行了改造,其主要措施是沿地洼地带挖排水沟,进行排水,降低地下水位,改良土壤的物理性质。但经过一段时间应用后,部分排水沟受到破坏,造成水土流失。为此,分析了黑龙江省中低产田改造治理过程中开挖排水沟产生水土流失的原因,并在此基础上提出了切实可行的防治措施。     

Controlled drainage — effects on drain outflow and water quality

A field experimental project was set up in southern Sweden to assess the effects of controlled drainage on hydrology and environment. Controlled drainage makes it possible to vary the drainage intensity with the variation in drainage requirement during season by controlling the height of a riser in the drain outlet and thus to a certain degree control the amount of outflow of solutes via the drainage system. During periods with low drainage demand, the riser in the drain outlet can be raised and the groundwater level in field will rise up to the level of the riser before the discharge takes place. Three plots, each with an area of 0.2 ha (40 m×50 m) were installed on a loamy sand. One plot was drained by conventional subsurface drainage (CD) and two plots were drained by controlled drainage (CWT). The plots contained four lateral drain tubes, at 10 m spacing and placed at 1 m depth. Each plot was isolated by a double layer of plastic sheeting placed in the back-filled trenches to a depth of 1.6 m to prevent lateral leakage and subsurface interactions. Measurements of precipitation, drain outflow and soil and air temperatures were carried out hourly. Groundwater levels were measured and samples of drain outflow were collected twice a month for nitrogen and phosphorous analyses. Mineral nitrogen contents in soil were measured three times a year.Controlled drainage had a significant hydrological and environmental effect during the 2 years of measurement (1996–1998). Compared with CD, the total drain outflow from CWT was 79% less in Year 1 and 94% in Year 2. The total reduction in nitrate losses with CWT corresponded to the reduced outflow rates. Compared with CD, the total amounts of nitrate in drain outflow were 78% less in Year 1 and 94% in Year 2. The highest concentrations of nitrate were measured at the time of the largest outflow rates. The phosphorous losses were 58% less for CWT as compared to the CD values in Year 1 and 85% less in Year 2. The reduction in nitrogen content in the soil profile during the winter season was 60–70% less in CWT than in CD.     

Shallow subsurface drainage in an irrigated vertisol with a perched water table

Waterlogging and salinity are reducing the productivity of irrigated agriculture on clay soils in south east Australia. We compared five drainage treatments: (1) undrained control (Control); (2) mole drains (Mole); (3) mole drains formed beneath gypsum-enriched slots (GES) (Mole + GES); (4) shallow pipe drains installed beneath GES (Shallow Pipe); (5) deep pipe drains (Deep Pipe). The experiment was set out on a vertisol and our measurements were made during the growth of an irrigated onion crop.

Over the 3 months before the spring irrigations commenced, the perched water table on the Control was less than 400 mm below the soil surface for 27% of the time, whereas the shallow drainage treatments (Treatments 2, 3 and 4) reduced this time to less than 4%. During the irrigation season, the perched water table on the Mole + GES treatment rose above 400 mm for 3% of the time. The perched water table on the Mole treatment was above 400 mm for 14% of the time, compared with 19% of the time on the Control. The Deep Pipes were less effective in reducing the depth to the perched water table, both before and during the irrigation period.

Mole drains increased the gas-filled porosity above the drains. However, the gas-filled porosity remained below reported levels for optimum root growth. Although the drains effectively drained excess water, and lowered the water table, the hydraulic gradient was insufficient to remove all of water from the macropores. Gypsum enriched slots above the mole drains increased the gas-filled porosity in the slots but the drainable porosity in the undisturbed soil appeared to be inadequate for optimum root growth, even though some drainage occurred near the slots.

Discharge from the shallow drainage treatments averaged 58 mm for each irrigation, and was considerably more than the amount required to drain the macropores. The mole channels were in reasonably good condition at the end of the irrigation season, with at least 70% of the cross-sectional area of the channel open.

Shallow subsurface drains increased onion yield by about 38%. For each day the water table was above 400 mm, the yield declined by 0.23 tonnes per hectare. Farmer adoption of shallow subsurface drainage will depend on the long-term economic benefits (influenced by the longevity of the mole channels and yields response) and the need to develop more sustainable management practices.