TN-6 : Estimation of seepage from canal using tracer technique

Abstract

Loss of water due to seepage from irrigation canals constitutes a substantial percentage of the total usable water. By the time the water reaches the field, it has been estimated that the seepage losses are of the order of 45 percent of the water supplied at the head of the canal. Of the various factors which influence seepage loss from a canal, the most important are the boundary conditions of the flow domain and permeability of the medium. Only after a correct assessment of the coefficient of permeability and boundary conditions, the seepage losses can be estimated either by numerical method or by analytical technique. Needless to say that it is very difficult to determine the in situ coefficient of permeability of the porous medium. In order to avoid the difficult task of estimating the coefficient of permeability and the prevailing subsurface boundary conditions experimental techniques like ponding method and inflow-outflow method have been used for estimation of seepage losses. In recent years, tracer technique has also been used to estimate seepage loss from water bodies because of its comparatively easy operation in respect to other experimental methods.

The seepage velocities at various sections of Deoband canal, a branch of Upper Ganga Canal, have been found by U.P. Irrigation Research Institute using nuclear tracer technique. These seepage velocities have been used subsequently in a formula derived by U.P.I.R.I. to estimate the seepage losses from the canal. In a study carried out at NIH, the U.P.I.R.I. formula used for estimating the seepage loss from tracer technique test data, has been rederived starting from the fundamentals. The losses estimated by the rederived formula using tracer technique data are found to agree with see-page losses measured by other experimental methods. Conformal mapping analysis has also been carried out to determine seepage loss from the canal.

From the analysis of seepage, it is found that the losses are only taking place through bed of the canal and the canal banks are relatively impervious. The deviation of the observed phreatic line from the locus of the phreatic line obtained analytically for a homogeneous flow domain justifies the above statement.