PERFORMANCE OF UNLINED AND LINED MOLE DRAINS IN A SALINECLAY LOAM FIELD

Sommerfeldt, T. G.Eng. 26: 1-5.

T. G. Sommerfeldt

Research Station, Agriculture Canada, Lethbridge, Alberta T1J4B1

Received 2 September 1983, accepted 11 January 1984

1984. Performance of unlined and lined mole drains in a saline clay loam field. Can. Agric.

The performance of unlined andlined mole drains was studied in a saline, poorly drained, clay loam soil in southernAlberta. Both mole drains were operative throughout the 10-yr study and were effective in lowering the water tableafter each flooding. However, discharge from the unlined moles averaged 24% of that from the lined moles. Drainspacings (5.0 m for theunlined, 7.6 m for the lined moles) were satisfactory foreffectively lowering the water table,and were in satisfactory agreement with the spacing calculated at 29 h, using hydraulic conductivity (K) from draindischarge. The Kvalue obtained bythe shallow well pump-in method (mainly horizontal flow) was 0.02 cm/h. By thedrain discharge method (horizontal and vertical flow) K was 0.07 and 0.22 cm/h for the unlined and lined mole drainareas, respectively, and by the water table recession method (horizontal and vertical flow) it was 0.17 and 0.28 cm/hfor the unlined and lined mole-drained areas. These data indicate vertical drainage was considerably greater than horizontal drainage, and that there was natural butslow internal horizontal drainage, possibly along thetill-bedrock contactplane. The function of thedrain after a flood was to aid the slow internal drainage in rapidly lowering the water tableto a depth at which extended waterlogging was nota problem. Thereafter, internal drainage andplant usewas sufficientfor continued lowering of the water table.

Reclamation of waterlogged saline soilswith mole drainage is uncommon. Yet,where conditions are suitable, mole drainage is less expensive to install than otherkinds of subsurface drainage. One reasonfor not using mole drainage may be theuncertainty of its life expectancy. In Britain, where mole drains are used on wetlands, their durability varies from days todecades with an average of 10 to 15 yr(Nicholson 1946). Soil texture and soilmoisture at time of installation are criticalfor stable mole drains (Childs 1942;Raadsma 1974; Sommerfeldt 1983;Spooret al. 1982a,b) and mole stability can beimproved by the installation of plastic linings (Dylla et al. 1963; Fouss and Donnan1962).

A field experiment was conducted in apoorly drained, saline, clay loam soil toevaluate the use of mole drainage and todetermine the relative performance of unlined and lined mole drains.

MATERIALS AND METHODS

Site Description and PreparationThe site, 100 x 200 m, was on a dom-

inantly saline, poorly drained, Brown Re-gosolic fluvial soil over till near Taber,Alberta. Saline-sodicsoil was also presentat the site. Texture of the fluvial mantle,0.6 m deep, and of the surface 0.6 m oftill was generally clay loam. Texture ofthe till below this depth was medium fine,but variedfrom very fine sandy clay loamtosiltyclayloamandclay loam, withgypsum crystals throughout. A tongue of bedrock at 2-m depth extended under about20% of the area (Fig. 1); elsewhere, there

were fragments of weathered bedrock at2.5 to 3.0-m depth, indicating bedrockwas near 3-m depth throughout.

The area wasformedto zerogradealongthe naturalcontours, witha 1%downslopefor minimum soil disturbance. North ofthe site, an open drain (item a, Fig. 1) wasconstructed into which the mole drains

discharged through galvanized tubing (90-mm diam). Mole drains (92-mm diam.)were installed in 1968 at 0.7-m depth, using materials and methods reported byFouss and Donnan (1962). Perforated 20-mil PVC plastic, notched on the edges,was used in the lined drains. As the liningwas fed through the mole plow, it foldedto form a tube and the notcheson the edgesinterlocked to retain the tube shape. Threedrains of a kind were installed in a blockwhich ran in a north-south direction frombottom to top of the slope. There were fourblocks each of unlined (U) and lined (L)drains randomly paired across the site, andthree check blocks (C) with no drains, oneon each end and one in the center of thesite. The C and U blocks were 15 m wideand the L blocks were 22.8 m wide. Between drains, spacing was 5 m for the Umoles and 7.6 m for the L moles (spacingsbased on past experience, E. Rapp, pers.commun.). Border dykes were constructed perpendicular to the contours at7.6-m intervals acrossthe site, for gravityirrigation. A tailwater drain (item b, Fig.1) was constructed at the north of the siteto collect runoff. A supply ditch across thesouth of the site (item c, Fig. 1) providedthe irrigation water. In thebankof the supply ditch, spiles (8 x 8 cm square with

CANADIAN AGRICULTURAL ENGINEERING, VOL. 26, NO. 1, SUMMER 1984

control gates) were installed, one to supply water to each border strip. The supplyditch provided sufficient capacity to irrigate the whole site in one setting. In boththe supply and runoff catchment ditches,non-submerged Parshall flumes with waterlevel recorders were installed to measurethe water delivered to and lost as runoff

from the site. Across the middle of the site,a row of 43 groundwater observationwells, of 2-cm perforated iron pipe, wasinstalled (Fig. 1) to monitor groundwaterdepth at prescribed times, using a pneumatic sounder.

The site was developed in 1968, seededto barley and flooded once in 1969. In1970, after it was seeded to tall wheatgrass(Agropyron elongatum) with a barley{Hordeum vulgare) nurse crop, intensivereclamation began.

Reclamation Procedures and DrainPerformance

The groundwater level was raised toground surface by flooding; groundwaterwas then drained away while the drain discharge and water table recession weremonitored. This procedure was carried outonce in 1969, and at 2- to 3-wk intervalsduring the irrigation season (5-7 times)from 1970to 1979. To saturate the profile,the site was flooded from 5 to 6 h on oneday to raise the water table to ground surface and again for about 3 h on the secondday. It was assumed that the soil was fullysaturated after the second flooding. Flooding was terminated each day when the tail-water drain, which had limited capacity,became filled to near-capacity. The net

%

-n • — " _^ )

1 94 A '

i

i

ii

IIi

•i

i

•

i

4

1

—I

: ci I Li Ui L2 U2

•

! c2 ;

OA. 7 '

u3 L3

i

i

i

ii

i

i•

t

•

i

t

i

Ua

1

I

i

L4 c3 |

: * 1• • • • • • • • * > • > • » • • 4 .' *0

• • f

La

i

> •

©ji

i

• • • •

ir^n

-t- 25.0 -*"1

!-|- 25.3 -!-

-

l_11

i

i

UPVl

i

i

i

i

t

i

i

i

-»• -»-»

SCALE:

—r—10

-1

20 meters

-»•

LEGEND

Drains

•••> Open drain

-»• Tailwater drain

Water supply ditch

Observation well

Observation well (recorder)

Parshall flume and recorder

Contour

Bedrock at 2m depth

-»

G

I24.4

Figure 1. Site plan showing check (C), unlined (U), and lined (L) blockswith locations of the shallow bedrock, mole drains, observation wells,open drain (a), tailwater drain (b), and supply ditch (c).

amount of water applied (total minus runoff) over the 2 days was 12-15 cm. Theaverage annual net application was 77 cm

"I

^ 0.40-yCDen

I 0.30-

LU

0.20-

< .

TO 0.10-

00

0.00

Figure 2. Average annual peak dischargefrom the lined and unlined mole

drains.

plus an annual average of 16.5 cm precipitation during the flooding season.

Drain discharge was measured with acalibrated vessel and stopwatch at termination of flooding on the second day andat 9, 18, 29, 46, 70, and 166 h later.Groundwater depths across the site weremeasured at the same times.

The grass growing on the site was cutwith a flail mower periodically throughoutthe growing season and left to decomposeon the soil surface.

In September 1976, soil samples to 2-m depth were collected along the row ofobservation wells. The electrical conductivity of the saturation extracts from thesesamples was determined, using methodsof the U.S. Salinity Laboratory (Richards1954). The sample sites within a drainedblock were at 1 m on either side of a drain

and midway between the drains, and atthree equally spaced distances across thecheck blocks.

At termination of the experiment in1979, the hydraulic conductivity (K) of theleached soil near the edge of the plot areaand of the unleached soil just outside theplot area (about 10 m apart) was determined by shallow well (1.2-m depth)pump-in method (Bouwer and Jackson1974) at eight locations, two on each sideof the area. These results were comparedwith K values determined with water tablerecession data (Dieleman 1974) and discharge data at 9 and 18 h (U.S. Bureau ofReclamation (USBR) 1978). Optimumdrain spacingsto achieve a specific drainage rate were calculated, using the transient flow method (USBR 1978) and Kvalues from drain discharge data, and

CANADIAN AGRICULTURAL ENGINEERING, VOL. 26, NO. 1, SUMMER 1984

compared with those used in the experiment.

RESULTS

The mean annual rate of peak drain discharge increased from 1969 to a maximumin 1974 for both the unlined and linedmoles, except in 1971 (Fig. 2). Thereafter, the rate decreased annually to timeof termination of recording in 1979.Throughout the study, discharge from unlined moles was less than that from thelined, the overall average being 24% ofthat from the lined drains.

Drain discharge during a flooding-leaching cycle decreased exponentiallywith time until it stopped after about 70 h(Fig. 3). Initially, the rate at which the dischargeapproachedzero was greater for thelined than for the unlined drains. But, after46 h, the discharge from the unlined andlined drains was similar.

Water table recession rates, midway between drains, in the first 9-h interval weregreatest in the blocks drained with linedmoles (Table I). Because of this, the watertable depths (y) in these blocks exceededthose in the unlined moles for the first 46

h. At that time, water table depths werewithin4 cm of drain depths (70 cm). Afterthe first 9-h interval, the rate of water tablerecession was greater from the unlineddrains than from the lined, which is attributed to the differences in head and drain

spacings. Noteworthy, though, is the rateof water table recession in the checkblocks. Relief for this water was attributed

to horizontal flow, especially along theplane of the weathered, fragmented bedrock at the till contact. The difference be

tween water table recession rates of the

check blocks and the blocks with unlined

and lined mole drains is attributed to the

drains.

If one assumes that drainage of this soilis primarily perpendicular to the restrictive layer (bedrock), K is a constantthroughout, and the head (H) is measuredfrom the bedrock (datum), then one cancalculate K for the check blocks after the

method of Klute (1965). The area of the

1400 -

10 20 30 40 50 60

Unlined

70

400 -

200

04

\1974\

\\

J 979 V

1970

10 20 30 40 50 60 70

TIME (h)

Figure3. Drain discharge forlined andunlined mole drains from afterflooding stopped (0 time)to after drainage stopped, in the years 1970, 1974, and 1979.

standpipe and cross-sectional area of thecore would be the same. Thus, Klute'sequation reduces to

K = {lit) In HJH2

where / is the length of saturated core (soil)which varies with time (t), as do //, andH2 which are the hydraulic heads at time1 and 2 (errors due to unsaturated flow andmatric suction are recognized).

Using this approach, the internal K ofthis soil, as determined from the watertable recession data of the check blocks

and a specific yield of 0.038 (= linedmole drain discharge - unit area/watertable recession), was 0.02 cm/h over thefirst four time intervals, i.e., to 46 h.(Evapotranspiration was not includedsince in the first few hours after floodingeverything was saturated, including all theorganic debris on the soil, and losses

through evapotranspiration should havebeen minimal.)

The average K values of the soil as determined by water table recession in thedrained plots, which measured vertical,radial, and horizontal (to bedrock) flows,were 0.17 and 0.28 cm/h for the unlined

and lined mole drained blocks. By thedrain discharge method at 9 h, whichmeasured vertical, radial, and horizontal(to effective drain depth) flows, the K values were 0.07 and 0.22 cm/h for unlined

and lined moles, respectively (0.05 and0.12 cm/h at 18 h). By the shallow wellpump-in method (inside and outside theplot area), which primarily measured horizontal flow, K was 0.02 cm/h, the sameas that estimated from the water tablerecession in the check blocks.

The calculated hypothetical drain spacings for this soil using K values from lined

TABLE I. MEAN WATER TABLE RECESSION RATES MIDWAY BETWEEN DRAINS (FOR ALL YEARS) AFTER FLOODING WAS STOPPED(TIME 0)

Check Unlined Lined

Depth to Difference Recession Depth to Difference Recession Depth to Difference RecessionTime, t water, y in depth, A y rate, A ylt water, y in depth, A y rate, A ylt water, y in depth, A y rate, A ylt

(h) (cm) (cm) (cm/h) (cm) (cm) (cm/h) (cm) (cm) (cm/h)

0 0 0 09 9.75 9.75 1.08 12.19 12.19 1.35 19.51 19.51 2.18

18 15.85 6.10 0.67 26.52 14.33 1.59 32.61 13.11 1.4529 24.99 9.14 0.82 37.19 10.67 0.96 41.45 8.84 0.8046 38.40 13.41 0.79 53.64 16.46 0.93 55.17 13.72 0.8170 48.46 10.06 0.41 67.36 13.72 0.57 66.45 11.28 0.47

166 65.23 16.76 0.17 86.87 19.51 0.19 83.82 17.37 0.17

CANADIAN AGRICULTURAL ENGINEERING, VOL. 26, NO. 1, SUMMER 1984

mole drain discharge data (at 9 h) were 6.0m, for a water table recession of 35 cm(half of drain depth) in 24 h, or 7.5 m fora recession of 41.45 cm in 29 h as ob

served (Table I). With K values from theunlined moles, the distance became 3.4 mfor a water table recession of 35 cm in 24

h, or 4.0 m for a recession of 37.19 cm in29 h as observed. Thus, at 29 h the calculated spacing for the lined drains wasessentially the same as that of the installeddrains, while that of the unlined drains was0.8 times that of the installed.

The electrical conductivity results of1976 for the surface 60 cm of soil (Fig. 4)show a decreasing salinity gradient fromthe center of Cx to the ^ block. Similargradients are seen from mid C2 into boththe U2 and U3 blocks. These results indicate that the radii of influence of the drains

extended into the check plots in somecases, suggesting that the drain spacingscould have been wider.

The cost of installing the drains in 1968was about $0.10/m for the unlined and$0.70/m for the lined. Freight was a majorportion of the additional $0.60/m cost forthe plastic liner. Also, the liner was experimental, which inflated its cost. Thecost of plastic was more than double thatquoted by Fouss and Donnan (1962) of$0.24/m. These material costs exceedthose for corrugated plastic drain tubing.But these costs could be reduced by manufacturing the liners near the place of use,as was done with the tubing.

DISCUSSION

Both the unlined and lined moles were

effective for draining this land to satisfactory depth within 24 to 48 h, though theunlined moles seemed to have restricted

capacity. The assumption is that the unlined drains were partially plugged by material falling from around the mole, possibly in the first year after installation,which caused the limited capacity. How-

12-

Q 4-|111

ever, these partially collapsed drains provided sufficient drainage to be effectivethroughout the length of the experiment.The results may not have been as favorable on a large-scale field installation.

Because the soil was saturated to groundlevel on two successive days it is assumedthat the profile was fully saturated whenobservations of groundwater recessioncommenced. Therefore, the observedrecessions are attributed to internal and

mole drainage. Lining the moles greatlyincreased their performance, namely indischarge, radius of influence, and apparent extended durability.

Flow among plots in the initial stages ofleaching was assumed to be insignificant.All plots were saturated to ground leveland the differences in groundwater depthsbetween plots were small. Also, K valuesfrom the shallow well pump-in test indicated horizontal flow would be small.

The variable K values are not in as much

disagreement with each other as their values suggest. The K values from the shallow well pump-in and water table recession (check block) methods are primarilyfor horizontal flow, assuming the bedrockis impermeable. The values from the draindischarge and water table recession (of thedrained blocks) methods are for vertical,radial, and horizontal flow, and the difference between the results is attributed to

the difference in depth of horizontal flowand experimental error. For the drain discharge method, depth is the effectivedepth of the drain whereas for the watertable recession, it is the depth to bedrock.

Though the study was terminated after10 yr of intensive leaching, the data suggest the drains would have functionedlonger. How much longer is unknown.

The drain spacings, 5 m for the unlinedand 7.6 for the lined drains, were narrowenough for the conditions of this study andare in satisfactory agreement with the calculated spacings. However, the electrical

o o o o o o

0 15 38 53 76 91 106 121 144 159 182 197

DISTANCE (m)

Figure 4. Soil salinity (EC) profile across the site at 0- to 60-cm depth in which the check (C),unlined drain (U) and lined drain (L) blocks are indicated as well as the locations ofthe unlined (x) and lined (0) drains. Drains in block U-l were not operating.

conductivity data indicate the spacingsmay have been a little too narrow.

The function of the drains was to aid the

slow internal drainage by rapidly loweringthe water table, after a flooding event, toa depth at which waterlogging was not aproblem (> 0.5 m). Thereafter, internaldrainage was sufficient and the water tablecontinued to recede slowly.

The conclusion from this study is thatmole drains were effective in draining thisland which had become salinized and wa

terlogged because of poor natural drainage. Lined moles were most effective.Land conditions were suitable at time of

drain installation, which is a requisite. Ifsoil conditions are favorable, there areeconomic advantages to mole drainage,because of the low cost of installation rel

ative to other kinds of subsurface drain

age. If the cost of lining material were reduced, the benefits from lining the molesmight be sufficient to justify the extra cost.Although lined moles might collapse overthe years, they would probably continueto function, albeit at reduced capacity.Further research is required to determinethe durability of mole drains, especiallylined, and to determine the soil conditionsunder which they perform satisfactorily.

ACKNOWLEDGMENT

Professor Egon Rapp started this experimentbefore joining the Agricultural Engineeringstaff at the University of Alberta, Edmonton.Appreciation is extended to N. Paziuk for hisdiligent services in overseeing and carrying outthe activities of this study.

REFERENCES

BOUWER, H. and R. J. JACKSON. 1974. Determining soil properties.Pages 611-672 in J. van Schilfegaarde ed.Drainage for agriculture. No. 17, Agron.Series, Am. Soc. Agron., Madison, Wis.

CHILDS,E.C. 1942. The mechanics of moledrainage. Emp. J. Exp. Agric. 10: 169-181.

DIELEMAN, P. J. 1974. Deriving soil hy-drological constants from field drainagetests. Pages 329-350 in Drainage principlesand practices. Vol. Ill, Surveys and Investigations. International Inst, for Land Re-clam, and Imp. Publication 16, Vol. III. P.O. Box 45, Wageningen, The Netherlands.

DYLLA, A. S., M. B. ROLLINS, and V. I.MEYERS. 1963. Experimental plasticdrains in Nevada. Nevada State Exp. Sta.,Reno, Nevada. Bulletin 228.

FOUSS, J. L. and W. W. DONNAN. 1962. Plastic-lined mole drains.

Agric. Eng. 43:512-515.KLUTE, A. 1965. Laboratory measurement

of hydraulic conductivity of saturated soil.Pages 210-221 in C. A. Black, ed. Methodsof soil analysis Part 1. Agron. 9. Am. Soc.Agron., Madison, Wis.

NICHOLSON, H.H. 1946. The principles of

CANADIAN AGRICULTURAL ENGINEERING, VOL. 26, NO. 1, SUMMER 1984

field drainage. Cambridge University Press, U.S. Dep. Agric. Handb. 60. Washington, failure of mole drainage channels. J. SoilCambridge, U.K. D.C. Sci. 33:411-425.

RAADSMA, S. 1974. Current drainage SOMMERFELDT, T. G. 1983. Soil mois- SPOOR, G., P. B. LEEDS-HARRISON, andpractices in flat areas of humid regions in ture and texture effects on mole drain sta- R. J. GODWIN. 1982b. Potential role ofEurope. Pages 115-143 in J. van Schilf- bility and onforce requirement for installa- soil density and clay minerology inassessinggaarde, ed. Drainage for agriculture. No. 17, tion —Alaboratory study. Can. Agric. Eng. the suitability of soils for mole drainage. J.Agron. Series, Am. Soc. Agron., Madison, 25: l^. Soil Sci. 33: 427-441.Wis. SPOOR, G., P. B. LEEDS-HARRISON, and U.S. BUREAU of RECLAMATION. 1978.

RICHARDS, L. A. (ed.) 1954. Diagnosis R. J. GODWIN. 1982a. Some fundamen- Drainage manual. U.S. Dept. of Interior,and improvement of saline and alkali soils. tal aspects of the formation, stability, and Bureau of Reclamation. Washington, D.C.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 26, NO. 1, SUMMER 1984