The Gadra Road Sub Branch, with a full supply discharge of 38.69m3/sec, is a canal of the Indira Gandhi canal project situated in desert western part of Rajasthan state, India. As part of this project, water from three Himalayan mountain range rivers – the Ravi, Vyas and Sutlej – has been diverted down to the desert to irrigate about 0.20Mha of land.

In 1989, soil scientist Dr. R.K.Katti recommended a cover of cohesive non-swelling soils on the bed and sides of the canal linings at a number of the canals which form the Indira Gandhi project. In 1997, for the portion reduced distance (R.D.) 39 (one R.D. is 304.8m) to R.D. 60, swelling pressure investigations were undertaken on disturbed soil samples with initial moisture content of zero. These were determined as 2.50 to 14.81kg/cm2, and therefore these soils were categorised as expansive. As per Indian Standard (IS) 9451:1994, cohesive non-swelling soil (CNS) cover should be provided over such soils below the canal lining. To determine whether a soil is expansive or not, a clay mineral test is normally undertaken, however, this was done completed in this case. As a result, the problem of expansive soils in the Gadra Road sub Branch was referred to the Indira Gandhi Canal Board for technical guidance in 1998.

Salient features of the Gadra Road Sub Branch are as below:

• Bed width – 5.64m.

• Full Supply Depth – 3.22m.

• Side slopes – 2:1.

• Full Supply Discharge – 38.69m3.

As the installation of CNS cover would cost an estimated Rs.28.7M (US$658,513), it was decided that detailed soil investigations should be undertaken in the portion RD 39 to 60.


The following soil investigations were completed to determine the expansiveness of the soils:

• Clay mineral test.

• Swelling pressure test.

• Mechanical analysis.

• Atterberg limits including shrinkage limit.

• Optimum moisture content and Proctor’s maximum dry density.

• Free swell index.

• Specific gravity.

• Triaxial shear test for cohesion and angle of internal friction.

• Total soluble solids.

• Sulphates.

• Carbonates.

• Organic matter.

• Dispersivity by Sherard’s PinHole test.

• Chemical analysis for pore water extract.

Clay Mineral tests for 16 soil samples were completed at the soil laboratory, TCS Division, Geological Survey of India, Western Region, Jaipur to discover the presence of the clay mineral Montmorillonite. The swelling pressure tests were undertaken at the soil laboratory of MBM Engineering College Jodhpur (Rajasthan: India). Soil investigations at S.N. 3 to 14 were conducted in the Material Testing Laboratory, Investigation Design & Research (Irrigation) Unit Jaipur (Rajasthan, India).

Results and discussion

Clay mineral test

Out of the clay mineral tests conducted on the 16 soil samples for the portion RD 39 to 60, Montmorillonite was only found as a minute trace in one sample at RD 44. Clay mineral Illite was found in small amounts or traces in some samples. These test results, as shown in Table 1, indicate that the soils are non-expansive.

Swelling pressure

It was discovered that swelling pressure tests conducted at MBM Engineering College, Jodhpur, in 1997 were not done correctly. The tests were repeated in 1998 using standard procedure and specifications. In the portion RD 39.5 to 47.5, swelling pressure of soil is more than 0.50Kg/cm2, and CNS treatment may be needed. A comparison of the soil tests conducted in 1997 and 1998 is shown in table 2.

Mechanical analysis

It is observed that clay content is more than 30% in the portion RD 42.50 to 47.50. The test results are available in Table 3

Atterberg limits

Liquid limit is higher than 50 at RD 45 to 47.5 and RD 55. Soils having liquid limit higher than 50 are considered expansive in nature. Soils at RD 42.5, 47.5 have plasticity index higher than 30 and can be categorised as expansive.

Optimum moisture content

Optimum moisture content at RD 42.5, 45, 47.5 and 60 is higher than the shrinkage limit in-dictating the possibility of crack formation in the soil on drying.

Free swell index

Free Swell Index varies from 13.63 to 52.00 i.e. lower than 100 indicating soils are non-expansive.

Triaxial shear test

As per Triaxial Shear Test at RD 45 and 47.5, the angle of internal friction is 13o and 5o indicating self stable side slopes as about 4:1 and 10:1. Depth of cutting plays an important role in side slopes.

Soil classification

Soils at RD 42.5, 45 and 47.5 are of CH group indicating soils are highly compressive and have high plasticity.


The carbonates are present in appreciable quantity in the soils.

Percentage sodium

As observed in Table 5, percentage sodium is higher than 60 at RD 39.5, 40, 45, 47.5 and 55 indicating soils may be dispersive in nature at these locations. However Sherard’s PinHole test results indicate soils are non-dispersive. Other test results do not show any adverse properties of the soil.

Looking at the above analysis, it was decided that there is no need to provide a CNS layer in the portion RD 47.5 to 60. With respect to the portion 39 to 47.5, it was observed that the canal is deeply cut and there is no possibility of any breach in the canal, even when run with full supply discharge.

The soil strata in this portion is relatively impervious and seepage in excess of the permissible limit would not occur even if the lining was not fully effective. Any delay in decision and construction work in this reach would cause delays in water supply and construction work in the downstream reaches. There would also be an estimated saving of US$658,513 by providing cover of locally available sandy soil instead of a CNS layer.

The alternative arrangement of constructing a new water supply channel would involve additional extra cost of about Rs7M (US$160,073) and its use would be limited to just one or two seasons.

There is considerable variation and contradiction in the laboratory test results. The GSI results show that there is no Montmorillonite constituent in the soil strata and therefore it is not bentonite material and will exhibit abnormal swelling property. The test results from MBM Eng. College laboratory have shown very different and varying values at different times and therefore cannot be fully depended on. The experimental lining completed without CNS layer in a 150m length in the reach has shown no distress. No cracks or swelling have been observed.

Taking into account all the above factors, it was decided that a CNS layer was not necessary and instead a cover of locally available sandy soil was provided over the existing bed and sides in the portion RD 39 to 60. Cement concrete lining was included as usual.


When planning the lining of canals, it would be useful to determine the properties of underlying soil. A soil can be termed as expansive only if the clay mineral Montmorillonite is present. To determine whether a soil is expansive or not, the simple preliminary test is free swell index test. A detailed investigation of the soil in the bed and sides of a canal is essential to determine the treatment to be provided.

Author Info:

The author is Y. C. Agrawal, Director (Retired), Minor irrigation Schemes, Department, Govern-ment of Rajasthan. Tel: +91(0) 141 2203508. Email


Table 1
Table 2
Table 3
Table 4
Table 5