Top Navigation Bar

Article Request Page ASABE Journal Article

Enhancing Subsurface Drainage to Control Salinity in Dryland Agriculture

H. Steppuhn, L. J. B. McArthur


Published in Applied Engineering in Agriculture 33(6): 819-824 (doi: 10.13031/aea.12252). Copyright 2017 American Society of Agricultural and Biological Engineers.


Submitted for review in January 2017 as manuscript number NRES 12252; approved for publication as part of the Advances in Drainage: Select Works from the 10th International Drainage Symposium Special Collection by the Natural Resources & Environmental SystemsCommunityof ASABE in August 2017.

Product and contractor names are provided for information only and do not confer endorsement.

The authors are Harold Steppuhn ASABE Member, Retired Research Scientist and L. J.Bruce McArthur, Associate Director, RDT, Swift Current Research and Development Centre, Science and Technology Branch, Agriculture & AgriFood Canada, Swift Current, Saskatchewan, Canada. Corresponding author: Harold Steppuhn, Swift Current Research and Development Centre, P.O. Box 3010, Swift Current, SK S9H 3X2, Canada; phone: 306-770-4426; e-mail: harold.steppuhn@agr.gc.ca.

Abstract. Controlling the physical processes of soil salinization involves lowering ground water levels, draining the vadose zones, and leaching excess salts from root zones. Plastic drain tubing strategically placed 1.5 to 1.8 m below the surface in semiarid lands can lower water tables and drain phreatic water, but irrigation is usually required to satisfactorily leach the offending salts. In non-irrigated drylands, the leaching process depends on natural precipitation, but the drier the climate, the greater the need for more leaching water. Possible practices which tap complementary water in conjunction with subsurface drainage include: (1) establishment of roughness barriers to trap wind-borne snow, and (2) pumping water from near-surface, ground water mounds. The mean electrical conductivity of saturated soil paste extracts sampled yearly from a semiarid site in Saskatchewan averaged 14.1 dS m-1 during the six years before the drainage was installed, 13.0 dS m-1 for two years just after drainage but before capturing blowing snow, and 9.6 dS m-1 for the six years following. The average barley grain harvested during the six years prior to drainage yielded 330 kg ha-1 and 2414 kg ha-1 after installation of the enhanced drainage system. In a follow-up sub-study, fall applications of 4.6 dS m-1 mounded ground water from a shallow well fitted with a solar-powered pump within a drainage system preceded spring seeding of alfalfa. Enhanced drainage improved mean seedling emergence from 20% to 79%. Every 28 mm of ground water applied, up to 2273 mm, increased alfalfa emergence by 1%.

Keywords.Agricultural drainage, Plant emergence, Pre-seeding irrigation, Solar-powered pumping, Soil reclamation, Soil salinity, Windbreaks.

Salinity has plagued agriculture in arid and semiarid climates for thousands of years. The geologic sediments underlying arid and semiarid soils typically contribute offending solutes (mostly sulfates, chlorides, and bi-carbonates) to the waters which move through them (Black, 1968). Whenever near-surface waters containing dissolved salts encounter drainage impediments in soils or substrata, a possibility exists for the solutes to concentrate. The installation of engineered subsurface drainage systems is recommended for lowering water tables, removing dissolved salts, and countering salinization by discharging excess water from root and vadose zones (Rhoades, 1974). Subsurface drainage of irrigated saline lands usually lowers water tables and reduces soil salinity (Rapp, 1968). But, a drainage installation in semiarid climates without the land being irrigated usually does not change salinity (Buckland and Hendry, 1992). However, after many years, non-irrigated drainage systems at some sites have decreased the electrical conductivity of soil solutions in the top 0.3 m of soil (Vander Pluym et al., 1985).

Black and Siddoway (1971) suggest that tall wheatgrass, Thinopyrum ponticum (Podp.) Lui & Wang, can serve as windbreaks on the Great Plains of North America. Tall wheatgrass, a salinity-tolerant, perennial bunchgrass, averages 1.2 m tall when placed in double-row wind-barriers spaced 15.2 m apart across agricultural fields. This grass shelterbelt system reduces ground wind speeds, deepens snow cover, conserves water, and increases and/or stabilizes crop production. Soil water enrichments are attributed to greater snow catch in the sheltered area compared to the open field and to reductions in evapotranspiration (Siddoway, 1970). Steppuhn and Waddington (1996) extended the merits of grass windbreaks to produce alfalfa crops after Miller et al. (1981) recommended growing alfalfa as a reclamation crop for watersheds with saline soils in recharge areas and wherever the crop would grow. The objectives of this study were to: (1) lower water table levels, (2) enhance snow cover for meltwater leaching of salts, (3) reduce root-zone salinity, (4) increase agricultural production, and, (5) maximize alfalfa seedling emergence after a one-time, pre-seeding irrigation in a semiarid climate using tall wheatgrass windbreaks combined with a subsurface drainage system.

Study Site

This study was conducted across a 23 ha site located 3 km southwest of Swift Current, Saskatchewan (50.3° N, 107.8° W). The mean annual precipitation at a nearby weather station equalled 359 mm, with up to one-third falling as snow. The growing season (May, June, July, August) precipitation and Class A pan evaporation averaged 210 and 883 mm, respectively. Two 1-ha parcels of severely salinized land, aligned north and south, characterized the site (fig. 1). The parcels form part of a slightly undulating plain sloping 3.4% to the southeast (fig. 2).

Figure 1. View northward across the salinity study site near Swift Current, SK, showing the South and North Parcels. (A saline, intermittent watercourse also shows northeast of the North Parcel.)

This site rests on ±210 m of the Cretaceous Bearpaw (Pierre) Formation beneath glacial till (Christiansen, 1959). Originating as marine sediment, this bedrock formation consists of dark grey, plastic montmorillonite shale which effectively impedes vertical drainage. The first of two post-Pleistocene continental glaciers deposited a plastic clay loam to sandy clay loam till (the Wymark Till) on top of the Bearpaw bedrock. As the Wymark Glacier retreated, it discharged proglacial sediment containing lacustrine marls, silts, and clays and fluvial sands and gravels leaving stratified drift covering the till. The second glacier (the Aikins) subsequently crossed the site, completely eroding the Wymark Till and its stratified drift on the down-hill (east) side of the site’s visibly saline parcels, before depositing the Aikins Till. The second till consists of plastic clay loam and completely covers the entire site to a thickness between 1.5 and 3 m. The Aikins glaciation was followed by a regional deposition of loess,estimated to have been 0.4 m thick at this site. The resulting soil of the plain is mapped as a degraded,salinized Swinton silt loam (Ayres et al., 1985) and is classified as a Salic Aridic-to-Aquic Haplo Boroll.

surface and subsurface geology, geomorphology, slope, aspect, and soils of the two salinized land parcels are hydrologically similar. The saturated hydraulic conductivity at the site measured by slug test (Bouwer and Rice, 1976) averaged 4.5 × 10-8 m s-1 for the Aikins Till and 4.7 × 10-7 m s-1 for the coarser Wymark Till (Helgason, 2000). The root-zone salinity at this site resulted from a “textural dam,” where the shallower bedrock and finer soil texture of the downhill east side of both land parcels caused ground water to accumulate in the vadose zone (Steppuhn, 2006).

Study Methods

In 1985, a grid of soil sampling points and wells spaced 7.6 m bi-directionally across the “textural dam” was established in each severely-salinized parcel (fig. 2). A perforated, plastic pipe (30 mm in diameter) within a surface-sealed hole augured down to a point just above bedrock at each location served as the observation well for weekly depth-to-water-table measurements. Every October or November from 1985 through 1996, a cylindrical core of soil, 50 mm in diameter, was removed from the upper 900 mm of the root zone within a 2 m radius of 10 selected sampling points per parcel. The cores were sectioned into 150 mm depth increments, and the electrical conductivity of the saturated soil paste extracts (ECe) determined for each increment following Rhoades (1982). The sodium adsorption ratios were not calculated, as sodicity is characteristically countered by the soil electrolyte concentrations typical of Swinton silt loam soils (Curtin et al., 1994).

In August and September of 1990, plastic drain-tubing covered with a polyester filter was buried on grade within 1.5 and 1.8 m of the surface of the South Parcel using a laser-controlled trencher operated by a commercial contractor (TC Farm West Drainage); the North Parcel was left untreated (fig. 2). The drain system consisted of north-south oriented, 100 mm, louvered plastic tubing spaced on 15.2-m centers feeding an east-west, non-perforated, 150-mm collector conduit. Drain lines were placed approximately 5 m west of the water wells and soil sampling points. The flow-rate of the subsurface drainage discharged from the treated parcel was measured with a sharp-crested, 15° V-notched weir using a Type-F Stevens float-cable, water-level recorder (volumetrically rated) monitoring stage with time.

Tall wheatgrass (variety: ‘Orbit’) was seeded as a windbreak in October 1991 and again in May of 1992 resulting in two side-by-side grass rows, 150 mm apart, across the South Parcel, reaching 0.7 m in height by November 1992 and 1.2 m the following spring. These grass windbreaks, grown parallel to and 3.8 m west of each drain line, effectively trapped snow beginning in winter 1992-1993.

We surveyed the accumulated snow cover on each parcel at or near peak accumulation each year beginning in January 1993, except during the snow-limiting winter of 1998. Each survey included 30 observations of snow cover depth and three to five specific gravity measurements of vertical snow cores on each parcel. These observations were used to calculate the mean snow cover water equivalent for each parcel according to Steppuhn (1976).

The mean depth-to-water-table and ECe values from the South Parcel obtained during the pre-treatment period (1985-1990) were regressed with those from the North Parcel using linear statistical models resulting in least-significant-differences (lsd) (SAS, 1995). The measured water table depths and ECe values were determined for the 0-450, 450-900, and 0-900 mm soil zones to provide the most reliable predictor for estimating the lsd values for these variables that would have occurred in the South Parcel had the subsurface drainage not been installed.

During the study, Bonanza barley was seeded across the entire site every year. Hand-harvested plant samples from each land parcel resulted in square-meter grain yields for comparisons before and after treatments. The samples were threshed and the grain oven-dried and weighed.

By 1998 and upon completion of the comparisons between the treated and untreated parcels, an engineered, subsurface drainage system was also installed in the North Parcel. The ground water mound under this parcel ranged within 2 m of the surface. This second drainage system included a 3.7 m deep well, dug with a backhoe, on the east edge of the Wymark Till (at 144.0 m North and 152.5 m East, fig. 2). Could the water from a subsurface drainage system be used for a one-time leaching of the offending salts sufficiently to establish alfalfa where severe salinity now exists? Alfalfa is the crop of choice in controlling soil salinization in dryland agriculture. In August 1999, an area within the severely saline parcel 15 m east of the well was disked, harrowed, and surveyed to delineate twelve 2.44 × 0.9 m separate plots (fig. 3). The 12 plots were randomized and six irrigated with in situ ground water pumped from the North Parcel well. The remaining six plots received no irrigation water. A direct-current, submersible pump, capable of lifting 15 L/min against a 4 m water head, was powered by a 75 W Siemens solar panel. The pumped water averaged 4.6 dS m-1 within a range of ±0.1 dS m-1 during the pumping period from 15 September to 23 October 1999; neither rain nor snow fell during this irrigation period. Twelve emitters spaced 305 mm apart along two parallel drip-lines each released a pressure-compensating 2.3 L/h. Water volumes per plot delivered to Plots 1 through 12 totaled 1181, 0, 1855, 0, 2193, 2530, 1012, 0, 1349, 0, 0, and 0 mm3 mm-2, respectively. After irrigation in the fall, the irrigation lines were removed and the plots left undisturbed over winter until seeded on 9 May 2000. Two rows, spaced 305 mm apart, of ‘Beaver’ alfalfa were carefully seeded per plot with a no-till forage plot seeder. Alfalfa plant emergence was measured on 7 July 2000. Measurements were obtained along a 1-m length of each sampled alfalfa seed-row. Each 20 mm segment along the length of the measurement meter (50 segments) was marked if it contained at least one emerged plant; 50 marks along the meter length inferred 100% emergence.

Figure 3. Surface view of the North Parcel experimental plots, water-supply well, water delivery and routing, and alfalfa seeding.

Results and Discussion

The annual precipitation during the six pre-treatment years exceeded the long-term mean (359 mm) in two years out of six in comparison to five out of the eight during the post-treatment years. Precipitation volumes averaged less before the drainage system was installed (316 mm) than after installation (412 mm). The total annual volumes per year of subsurface water discharged from the South Parcel drainage system equalled 912, 336, 1577, 878, 1936, 2490, and 1296 m3 in 1991 through 1997, respectively.

Once established, the tall wheatgrass windbreaks accumulated more snow within the wind-sheltered (South) parcel than across the open (North) parcel. Over the six surveys (1993-1999 excluding 1998 because of insufficient snow), the mean peak in areal water equivalent averaged 72.8 mm on the South Parcel and 25.5 mm on the North Parcel. Surface to water-table depth measurements (n=559) taken at least once weekly during the pre-treatment period (1986-1990) correlated closely between the two parcels (r2=0.887, standard error <0.006 m) (fig. 4). After the subsurface drainage system was installed in the South Parcel, but before the tall wheatgrass windbreaks became effective in 1993, the average depth-to-water (n=331) in the ten wells within each of the parcels equaled 1.74 and 1.30 m for the South and North, respectively. After the snow-trap grass windbreaks became effective (from 1993-1997), the South Parcel water level depths averaged 1.72 m and the North 1.07 m (n=261). The differences in the measured and predicted depths (based on the North Parcel measurements) averaged across the South Parcel for the drainage only and the drainage plus windbreak treatment periods equaled 0.478 and 0.693 m (lsd(0.01): ±0.038 and ±0.056 m), respectively.

Figure 4. Mean depth to static ground water level from soil surfaces plotted with time for the ten selected observation wells in the South Parcel (pre-treatment, engineered subsurface drainage only, and drainage plus grass windbreak periods) and the ten selected wells in the North Parcel (untreated control), near Swift Current, SK; depths of drain lines ranged from 1.5 to 1.8 m.

The soil-EC covariance between the North and South Parcels during the six pre-drainage years (1985-1990) showed significant correlations in the 0-450 mm (r2=0.95) and 450-900 mm (r2=0.74) soil zones but not when combined into the 0-900 m zone (r2=0.48). The regression slope for the 0-450 mm unit approached unity (0.958), and the intercept indicated that the South Parcel, on average, was more saline (by 2.3 dS m-1) than the North Parcel during the pre-treatment years.

Figure 5. Average ECe to a depth of 450 mm for each year from soil samples extracted in October or November of each year. The drain tubing in the South Parcel was installed during 1990.

The lsd at pa<0.05 for predicting the 0-450 mm soil ECe in the South Parcel based on the North Parcel measurements prior to drainage equaled ±0.66 dS m-1 (±0.86 dS m-1 at pa<0.01). This implies that the reductions in the 0- 450 mm root-zone salinity of the South Parcel relative to that for the North Parcel after the 1990 drainage and the 1992 grass windbreaks are significant (fig. 5). The average post-treatment difference between the measured and predicted South ECe equaled 6.46 dS m-1 and exceeded the 0.86 dS m-1 lsd by over seven-fold.

Barley grain harvests from the parcels during 1984-1990 (before treatments), yielded 330 kg ha-1 or less. After installation of the subsurface drainage system, grain yields averaged 2413.6 and 140.4 kg ha-1 from the South and North Parcels, respectively.Within five years after installing the drainage and adding snow management across the South Parcel, the aerial view of the site changed from figure 1 to figure 6.

Figure 6. View northward showing South Parcel after subsurface drainage and tall wheatgrass windbreak treatments, the untreated North Parcel, and the saline, intermittent watercourse at the salinity study site near Swift Current, SK.

In the one-time irrigation study, the alfalfa plants began to emerge just before 19 May 2000, ten days after seeding (Steppuhn, 2001). By 23 May, the rates of emergence observed in the irrigated plots visually exceeded those in the non-irrigated plots with differences increasing as the growing season progressed. On 7 July, mean emergence percentages significantly favored those plots pre-treated with irrigation: 78.8% over 19.7% (standard error = 12.0% and 6.99%), respectively. The objective of this one-year alfalfa sub-study was to demonstrate a possible value-added use of a ground water mound in a subsurface drainage system installed in a saline site under dryland agricultural production.

The total volume of in situ water applied to each irrigated plot varied from 1012 to 2530 mm3 mm-2. Together with the six non-irrigated plots, these volumes formed the independent variable for which a regression could determine any dependence of the alfalfa emergence on the pre-irrigation. The statistical results from the linear regression indicated that the pre-seeding irrigations explained 75% of the measured variation (r2=0.75) in alfalfa plant emergence (E in % per meter row length) according to an empirical relationship,

E = 35.6 (Vi) + 19.1 (2)

where Vi equals irrigated water volume in meters per square meter. A 100% emergence was reached with 2.273 m of applied water.

Conclusions

This study has led to five inferences: (1) that snow management can increase the water available to leach salinized root zones, (2) that engineered subsurface drainage can lower water table elevations, especially at the critical time just before spring seeding, (3) that a combination of subsurface drainage and grass windbreak technologies in semiarid, non-irrigated environments can result in a gradual reduction of salinity in the top 450 mm of soil, (4) that the combined drainage-windbreak treatment can increase barley grain yields 17-fold, and (5) that pre-seeding irrigation with mounded, in situ, ground water can increase the percentage of seeded alfalfa emergence.

Besides lowering water tables, the resulting advantages of installing engineered subsurface drainage systems in saline soils underlain by elevated ground water mounds include: dryer soil surfaces giving better ability to bear farm implement traffic; fuller utilization of applied fertilizers to grow crops; more efficient use of field time by farming across saline soil parcels rather than diverting around them; gaining more property value and land to use.

Three possible additional benefits (recommendations) derived from this study include:

  1. the installation of a subsurface drainage system in a dryland agricultural site within a semiarid climate can result in harvestable water for various uses; within the bounds of the water’s quality, these waters can serve livestock, establish woody windbreaks, irrigate gardens, or leach salts from saline soils;
  2. the establishment of alfalfa in saline soils within ground water discharge (seep) areas fitted with an engineered subsurface drainage well can lead to a common reclamation crop across the same field when alfalfa is also seeded in the field’s recharge areas; this reduces costs, simplifies cropping, contributes to greater productivity, and increases financial return;
  3. the reclamation of severely saline soil with waters from an engineered subsurface drainage system can be deferred until post-harvest but before the onset of winter; either barley or alfalfa can be seeded in the following spring.

The windbreak system in our study covers a limited area of 1.1 ha. If this system, because of its small size and upwind fetch, allowed extra measures of snow to accumulate in it, a larger windbreak system may not prove as beneficial. Also, if the subsurface water volumes or their dissolved salt concentrations discharging from the drainage system are large, additional disposal techniques may be necessary.

Acknowledgements

This study was conducted on the Smith Brothers Farm. On behalf of Agriculture and Agri-Food Canada, the authors gratefully acknowledge and thank Murray, Allan, Stuart, and Doug Smith for allowing the use of their land and their many contributions to this study. Thanks and appreciation are also extended to Don Sluth, Evan Powell, Darcy Schott, Tom Olfert, Gary Winkleman, Rod Ljunggren, Del Jenson, Ray Bunnell, Warren Helgason, Craig Gatzke, Gordon Evjen, Gerry Luciuk, Brook Harker, Dean Smith, Walt Nicholaichuk, Bob Zentner, Terry Speed, and many other contributors at the Semiarid Prairie Agricultural Research Centre and PFRA. In particularly, Ken Wall and Ken Deobald additionally deserve very special acknowledgment and thanks for their contributions, dedication, and insights associated with this study.

References

Ayers, K. W., Acton, D. F., & Ellis, J. G. (1985). The soils of the Swift Current map area 72J Saskatchewan, Saskatchewan Institute of Pedology, Publ. 481. Saskatoon: University of Saskatchewan Extension Division.

Black, A. L., & Siddoway, F. H. (1971). Tall wheatgrass barriers for soil erosion control and water conservation. JSWC, 26, 104-111.

Black, C. A. (1968). Soil-plant relationships (2nd. ed.). New York, NY: John Wiley & Sons.

Bouwer, H., & Rice, R. C. (1976). A slug test for determining hydraulic conductivity of unconfined aquifers with completely or partially penetrating wells. Water Resour. Res., 12(3), 423-428. https://doi.org/10.1029/WR012i003p00423

Buckland, G. D., & Hendry, M. J. (1992). Groundwater response and salt removal in a saline-seep soil in southern Alberta. Canadian Agric. Eng., 34(2), 125-134.

Christiansen, E. A. (1959). Glacial geology of the Swift Current area. Report No. 32 (Reprinted 1969). Saskatoon, Saskatchewan: Sedimentary Geology Div., Geological Sciences Branch, Dept. Minerals Resources.

Curtin, D., Steppuhn, H., & Selles, F. (1994). Clay dispersion in relation to sodicity, electrolyte concentration, and mechanical effects. SSSAJ, 58(3), 955-962. https://doi.org/10.2136/sssaj1994.03615995005800030045x

Helgason, W. D. (2000). Evaluation of subsurface drainage techniques used for dryland salinity control. MS thesis. Saskatoon: University of Saskatchewan, Division Environmental Engineering.

Miller, M. R., Brown, P. L., Donovan, J. J., Bergatino, R. N., Sonderegger, J. L., & Schmidt, F. A. (1981). Saline seep development and control in the North American Great Plains: Hydrogeological aspects. Agric. Water Manag., 4(1), 115-141. https://doi.org/10.1016/0378-3774(81)90047-0

Rapp, E. (1968). Performance of shallow subsurface drains in glacial till soils. Trans. ASAE, 11(2), 214-217. https://doi.org/10.13031/2013.39375

Rhoades, J. D. (1974). Drainage for salinity control. In J. van Schilfgaarde (Ed.), Drainage for agriculture (Ch. 16, pp. 433-461). Agronomy Monograph 17. Madison, WI: ASA.

Rhoades, J. D. (1982). Soluble salts. Part 2: Chemical and microbiological properties. In A. L. Page (Ed.), Methods of soil analysis (2nd. ed., pp. 167-179). Madison, WI: ASA, SSSA.

SAS. (1995). Statistical discovery software, JPM. Ver. 3.2.1. Cary, NC: SAS Institute, Inc.

Siddoway, F. H. (1970). Barriers for wind erosion control and water conservation. JSWC, 25(5), 180-184.

Steppuhn, H. (1976). Areal water equivalents for prairie snowcovers by centralized sampling. Proc. 44th Annual Western Snow Conf., (pp. 63-68). Calgary, Alberta: Pub. WSC 614-76. URL:sites/westernsnowconf.org/PDFs/1976

Steppuhn.pdf

Steppuhn, H. (2001). Pre-irrigation of a severely-saline soil with in situ water to establish dryland forages. Trans. ASAE, 44(6), 1543-1551. https://doi.org/10.13031/2013.7039

Steppuhn, H. (2006). Combining subsurface drainage and windbreaks to control dryland salinity. Canadian J. Soil Sci., 86(3), 555-563. https://doi.org/10.4141/S05-052

Steppuhn, H., & Waddington, J. (1996). Conserving water and increasing alfalfa production using a tall wheatgrass windbreak system. JSWC, 51(5), 439-445.

Vander Pluym, H. S., Livergood, C., & Healy, L. (1985). Development of agricultural subsurface design standards for Alberta. Final Report, Farming for the Future Project 83-2042. 83-2042. Lethbridge: Alberta Agriculture.