A method for evaluation of artificial recharge through percolation tanks using environmental chloride

A method for evaluation of artificial recharge through percolation tanks using environmental chloride

Sukhija, B S


About two-thirds of India, comprising Southern Peninsular and Western India, is underlain by hard rocks consisting of granites, gneisses, charnockites, basalts, etc. This land mass receives most of its rainfall during June to September. The pulsed nature of the rainfall, the low infiltration rate in soil, and the poor storage underneath result in a large runoff component. The net recharge to ground water in such environs is usually only a few percent of the average total rainfall (Sukhija, 1978; Sukhija et al., 1996). A continued increase in population has resulted in greater agricultural and industrial growth in recent times. This has only been possible by greater ground-water exploitation, leading to continued decline of the water table. Dug wells usually go dry in early summer, thus there is a need for artificial recharge. Of the various possible artificial recharge methods in Peninsular India, percolation tanks are currently the most popular. A percolation tank impounds the water of a monsoon stream behind a small earthen dam (Figure 1) with the hope that water collected in this tank during the rainy season will recharge the ground water until infiltration and evaporation exhaust the entire collection of the tank. (Figure 1 omitted) Typically, there are a number of wells on the downstream side of the percolation tank (Figure 1), which are expected to be recharged by the water that has infiltrated in the bed of the tank. The construction cost of a medium size percolation tank is about half a million rupees (about U.S. $15,000). Thousands of such tanks are constructed every year. As no reliable and practical method is available to evaluate their performance in recharging the ground water, the opinion has sometimes been expressed that the “percolation tanks” are merely “evaporation tanks.” In view of the large monetary investment being made, it is vital to develop a reliable, simple, and inexpensive method for evaluating the performance of these tanks.

In this paper, we describe the development of a method based on a series of environmental chloride measurements in the tank water. The results from a pilot experimental study encompassing two consecutive years are presented here.

Previously two methods have been utilized to evaluate the performance of the percolation tanks in India: the water balance method and the environmental stable isotope and/or radiotracer method. In the former approach, the water level in the tank is monitored and the changes are ascribed to the sum total of evaporation and percolation. Evaporation is either computed from a handful of meteorological data or estimated from measurements of an evaporimeter installed at a meteorological station. The percolation is obtained from the difference of change in water level and estimated evaporation. Naturally, the accuracy of estimation of percolation from the tank depends on the accuracy of the estimation of evaporation. This approach was adopted by Raju (1985) as well as Sharma (1985). They estimated that about 75-80 percent of water collected in tanks in the southern and western part of India percolates down to recharge ground water. The accurate evaluation of evaporation is, however, often problematic. If the water level in the tank falls daily by more than 2 cm, then the tank may be considered to be dominantly percolative since daily evaporation in winter months is considerably less than 1 cm. But in most tanks, the water level falls by ==1 cm daily, a condition requiring accurate determination of evaporation in order to assess percolation. Under the second approach, Nair et al. (1978) used the enrichment of deuterium and oxygen–18 in the tank water (due to evaporation and trace the enrichment of these isotopes in ground water) to delineate the area influenced by the percolation tanks. The relatively sophisticated and expensive methodology of environmental stable isotopes which, in general could be useful, the method in this specific study provided only qualitative answers to the problem. This was especially due to seepage from canals in the area which was also enriched in stable isotopes, thus making it difficult to differentiate between the two sources of recharge in the area. Another attempt (Nair et al., 1980) labeling the tank water with artificial tritium (100 curies) also did not provide quantitative information. In contrast to these methods, the present environmental chloride method is simple and inexpensive, yet reliable for the purpose required.

Environmental Chloride Method

Environmental choride is deposited on the land from the atmosphere by rainfall and dry fallout. Jacks (1980) made an attempt to estimate the dry fallout in southern India, about 200 km away from the coast, to be about 0.02 mg/cm sup 2 /yr. It is expected that the fallout will be lower in the further inland areas. In the case of the percolation tank, a contribution from dry fallout is expected to be completely negligible since the wet fallout is collected via runoff from a large catchment area while the post monsoon component of dry fallout occurs over a relatively small area occupied by the tank. In recent years there has been an increase in utilizing environmental chloride as a tracer for various geohydrological studies. This is due to two reasons: (1) chloride is a conservative tracer (Hem, 1970; 1985) and (2) it can be measured using a simple technique. In situations where there are no additional sources or sinks of chloride, the method can be effectively used.

Environmental chloride tracer has been previously used to estimate the natural ground-water recharge rate. This method has been applied in a semiarid region of Israel by Eriksson and Khunaksem (1969), in India by Sukhija and Rama (1973) and Allison and Hughes (1974, 1978) and Sharma and Hughes (1985) in Australia to estimate the natural ground-water recharge as well as that due to change in the land cover. Edmunds et al. (1990) used the method for past and present recharge evaluation in semiarid and arid terrains. Natural ground-water recharge was estimated in coastal aquifers of Pondicherry, India, by measuring the chloride present in soil profiles in repeated experiments, and the results were found to be comparable with those obtained with the injected tritium method (Sukhija et al., 1988). Sharma and Hughes (1985) have used chloride in the soil profile and in ground water to demonstrate the importance of the by-pass mechanism in soil-water movement, and Taniguchi and Sharma (1990) have used the tracer approach to estimate the ratio between mobile and total water content in a soil profile.

Sukhija and Reddy (1987) have demonstrated the usefulness of the chloride method to detect the presence of percolated water in wells situated downstream of a percolation tank. This was possible because the chloride concentration of the tank water soon after the rainy season was very low compared to that in the normal ground water. As a consequence, the low chloride concentration was observed in the wells just downstream of the tank influenced by its recharge. The method can qualitatively delineate the extent of influence of the tank in the wells located downstream.

Present Approach

Assuming that there are no sources or sinks of chloride in the percolation tank other than natural input from precipitation and runoff before impounding, and further that there is no loss of water by factors other than evaporation and percolation (seepage from the dam, if any, should also be accounted), the mass balance of chloride in the tank water can be used in estimating the percolated fraction of total volume of water. When the tank dries up during summer, some chloride stays back in the top soil. This may mix with the tank water when the tank fills again. Since we start our observations at the end of the rainy season, i.e., after three or four months of water collection, and compute the balance of chloride in tank water before the tank completely dries up, we expect this source to become unimportant. Further, it is assumed that the contribution of chloride by diffusion from soil to the tank water during the experimental period would be negligible, though it needs to be ascertained.

The total chloride content of the tank water at any time is estimated from the volume of tank water at that time and its chloride concentration which is measured regularly (for example, every week after the end of the rainy season). As there is no loss of chloride by evaporation, the chloride concentration of the tank water should increase with time because of evaporation. This generally happens, as there are long dry spells of several weeks after the monsoon. The percolating water, however, takes the dissolved chloride with it. Thus, by measuring the chloride concentration and the volume of water in the tank at different times (after the monsoon), a reliable estimate of the evaporation from the tank can be made. Assuming no seepage/leakage from the dam, the percolation fraction of the tank water can then be calculated.

The chloride balance in the tank water between time t sub 1 and t sub 2 can be written as follows:

V sub 1 C sub 1 = V sub 2 C sub 2 + (1 – f) (V sub 1 – V sub 2 ) C sub p (1)

where C sub p = SigmaC sub i sub i /SigmaV sub i = time weighted average concentration of chloride in percolated water.

(1 – f) = (V sub 1 C sub 1 – V sub 2 C sub 2 )/C sub p (V sub 1 – V sub 2 ) (2)

where V sub 1 = volume of water in the tank (after monsoon) at time t sub 1 ; C sub 1 = chloride concentration in the tank water at time t sub 1 ; V sub 2 = volume of water in the tank at time t sub 2 ; C sub 2 = chloride concentration in the tank water at time t sub 1 ; V sub 2 = volume of water in the tank at time t sub 2 ; V sub 1 – V sub 2 = water loss from the tank between t sub 1 and t sub 2 ; f(V sub 1 – V sub 2 ) = loss of water by evaporation; (1 – f)(V sub 1 – V sub 2 ) = loss of water by percolation; f = fractional loss by evaporation; and 1 – f = fractional loss by percolation.

The percolated fraction (recharge) can be calculated if V sub 1 and V sub 2 , and C sub 1 and C sub 2 are known. The weighted concentration C sub p of the percolated water is estimated from SigmaC sub i V sub i /SigmaV sub i , weighting done from V sub 1 to V sub 2 . It is assumed that the fractional percolation (and also the fractional evaporation loss) remains constant during the period (t sub 1 to t sub 2 ), i.e., when volume decreases from V sub 1 to V sub 2 .

Location of Tank and Field Measurements

A pilot tank, which was constructed around 1989, was selected in typical granite/gneisses type rocks near Singaram village, Yacharam mandal, Rangareddy district (around Hyderabad) in Andhra Pradesh, India (Figure 2) which is about 400 km away from the coast. (Figure 2 omitted) The water-covered area of the tank is about 15,000 sq. m, with about 170 m length and about 2.5 m height of tank bund, and with about 10,000 cu. m. of full capacity. In order to determine the volume of water in the tank at different stages, a detailed topographic survey of the tank bottom at a 5 m grid interval was carried out. A scale with 1 cm graduation was installed in the tank near the bund. A detailed (1 cm interval) stage vs. volume curve was drawn to determine the water quantity available at different times (Figure 3). (Figure 3 omitted) Measurement of daily water level in the tank was initiated in November 1992 (after the rainy season) and was continued until the tank was dry. Two sets of measurements during 1992-93 and 1993-94 are available for comparison. The water samples were collected for chloride measurement in the tank water on a weekly basis. Measurements of water level in the tank and in ground water have indicated the movement of water from tank to ground water.

Chloride Measurements

Chloride concentration in the water sample was measured by the mercuric thiocyanate and ferric nitrate (colorimetric) method using a spectrophotometer (Navada, 1982). The chloride ions react with mercuric thiocyanate to form soluble mercuric chloride, releasing the thiocyanate ions. In the presence of ferric ions a highly colored ferric thiocyanate complex is formed. The absorbance of this complex is proportional to original chloride concentration. Using this technique it has been possible to measure chloride concentration down to 1-2 ppm; concentrations in the range of 5-30 ppm could be measured with a precision better than 10%.

Results and Discussions

Measured water levels in the tank for about 4 1/2 months (end of November 1992 to April 1993) are plotted in Figure 4. (Figure 4 omitted) As can be seen from the figure, the decline of the water level (due to percolation and evaporation) is quite linear, but comprises two segments. Segment one is during the winter period (November to the middle of February) when reduction in water level is at the rate of about 0.58 cm/day while for the second segment, in summer (end of February and March), the rate is observed to be about 0.77 cm/day. Using the measured water levels, the volume of water present in the tank at different times was obtained from the stage-volume curve of Figure 3. Volume of water in the tank vs. time, and the measured weekly chloride concentration are shown in Figure 5. (Figure 5 omitted) Similar to the water level curve, for the first three months following impounding, the increase in the chloride concentration with time is approximately linear. Later, when the quantity of water in the tank becomes meager, the concentration increases many folds. A smooth curve is drawn through the available data for the period November 1992 to February 1993. Most of the recharge from the tank has taken place during this period. High chloride concentration is evident after 90 to 110 days. This is largely due to the fact that the volume of water in the tank has reduced considerably (==750 cu. m compared to about 3500 cu. m initially, and evaporation amounts to progressively larger volume fractions.

Table 1 shows the weekly chloride concentration along with the total quantity of water in the tank on that day. (Table 1 omitted) These two measured quantities form the basis of equation (1). Progressive four week percolation fractions are calculated and shown in Table 1. The four week interval was chosen so that the change in chloride concentration was sizable and accurately measurable. The method, therefore, is applicable only if dry spells last for more than a month (which happens to be the case in most parts of India, where dry spells last for months). The progressive percolation fraction shows time variation (Table 1), i.e., maximum percolation fraction during December-January, when the evaporation rate is expected to be minimum. Minimum percolation fraction is discernible in the month of February when the evaporation rates are higher.

The second half of Table 1 shows the results obtained using the conventional water balance technique. Here the percolated fraction is obtained from the loss of water level in the tank minus evaporation during the same period (evaporation loss is directly obtained from the evaporimeter data collected at Hyderabad Airport station of Indian Meteorological Department).

Based on the abvove methodology, average percolation fractions and percolation rate (average of seven weeks) were also worked out for 1993-94 and compared for the corresponding weekly average (Table 2). (Table 2 omitted)

The analysis of percolation data (Table 2) shows that the percolation rates, as determined using chloride method for 1993-94, show a falling trend very conspicuously when compared with the corresponding time period of 1992-93. Such a variation is not observed using the water balance method. The reduction in percolation rates can be expected due to siltation of the tank. Furthermore, unlike as in the case of the chloride method, the water balance method does not indicate any significant time variation (seasonal or annual). The water balance method must make use of questionable evaporation data (measured by evaporimeter at the meteorological observatory). The chloride method, on the other hand, evaluates evaporation and recharge using directly measured quantities with some valid assumptions. Thus, the latter can be considered more reliable.

As we can see from Tables 1 and 2, the average monthly percolation fraction of the tank volume is about 30-35% using the chloride method and about 50% using the water balance method.


Percolation tanks are being constructed extensively in India with the hope of recharging the ground water artificially. It is vital to develop appropriate methods to estimate recharge to ground water through these tanks. We demonstrated the use of environmental chloride to estimate the recharge through the tanks to ground water. The method developed is simple, inexpensive, and practical, and provides authentic measured recharge rates at various stages of silting (and desilting) of the percolation tanks.


The authors are grateful to Dr. H. K. Gupta, Director, NGRI, for his inspiration and kind permission to publish this paper. We are also thankful to the A. P. State Irrigation Department for allowing us to make use of their percolation tanks for this study, and Sri M. Dhanraju for his help in the field surveys. We are grateful to the anonymous reviewers for greatly improving the presentation of the paper.


Allison, G. B. and M. W. Hughes. 1974. Environmental tritium in the unsaturated zone; estimation of recharge to an unconfined aquifer. Proc. Symp. Isotope Techniques in Groundwater Hydrology. IAEA, Vienna. v. 1, pp. 52-72.

Allison, G. B. and M. W. Hughes. 1978. The use of environmental chloride and tritium to estimate total recharge to an unconfined aquifer. Aust. J. of Soil Res. v. 16, pp. 181-195.

Edmunds, M. W., W. G. Darling, and D. G. Kinniburgh. 1990. Solute profile techniques for recharge estimation in semi-arid and arid terrain. Groundwater Recharge, IAH. v. 8, pp. 257-270.

Eriksson, E. and V. Khunakasem. 1969. Chloride concentration in groundwater, recharge rate and rate of deposition of chloride in the Israel coastal plain. J. of Hydrology. v. 7, pp. 178-197.

Hem, J. D. 1970. Study and interpretation of the chemical characteristics of natural water (2nd ed.). U.S.G.S. Water Supply Paper 1473. 365 pp.

Hem, J. D. 1985. Study and interpretation of the chemical characteristics of natural water (3rd ed.). U.S.G.S. Water Supply Paper 2254. 263 pp.

Jacks, G. 1980. Hydrogeochemical studies. SIDA assisted groundwater project in Noyil, Ponnani and Amaravati River basins, Tamil Nadu & Kerala, India. Central Ground Water Board, Report 1:5. 38 pp.

Nair, A. R., A. S. Pendharkar, S. V. Navada, and S. M. Rao. 1979. Groundwater recharge studies in Maharastra–Development of isotope techniques and field studies. Isotope Hydrology 1978, IAEA. v. II, pp. 803-825.

Nair, A. R., S. K. Jain, S. M. Rao, and A. C. Eapen. 1980. Radio tracer technique to study the efficacy of Bangarwadi percolation tank. Proc. Workshop on Nuclear Techniques in Hydrology held at NGRI, Hyderabad, India. pp. 219-230.

Navada, S. V. 1982. Comparison of the analytical methods for analysing trace amounts of chloride. Inst. of Geol. Sci., Wallingford, Oxfordshire. Unpub. rep. WD/ST/82/8.

Raju, K.C.B. 1985. Recharge through percolation tanks and subsurface dykes, India. Proc. Semi. on Artificial Recharge of Groundwater held at Ahmadabad, India. pp. 12c-1.

Sharma, S. C. 1985. Artificial recharge in Gujarat state by injection, spreading through percolation tanks, check dams, Bandharas & other methods. Proc. Semi. on Artificial Recharge of Groundwater held at Ahmadabad, India. pp. 16c-1.

Sharma, M. L. and M. W. Hughes, 1985. Groundwater recharge estimation using chloride, deuterium and oxygen-18 profiles in the deep coastal sands of Western Australia. J. of Hydrology. v. 1, pp. 93-109.

Sukhija, B. S. and Rama. 1973. Evaluation of groundwater recharge in semi-arid regions of India using environmental tritium. Proc. of Ind. Acad. of Sci. v. LXXVII, sec. A, no. 6, pp. 279-292.

Sukhija, B. S. 1978. Groundwater recharge rates in semi-arid regions of India using environmental tritium. Proc. on Study and Management of Water Resources in Semi-Arid Regions of India held at Ahmedabad. Arid Zone Hydrology, ed. by S. K. Gupta and P. Sharma, 1979. pp. 103-116.

Sukhija, B. S. and D. V. Reddy, 1987. Study of artificial recharge through Potharam percolation tank by environmental chloride method. Geophy. Res. Bull. v. 25, no. 1, pp. 27-31.

Sukhija, B. S., D. V. Reddy, P. Nagabhushanam, and R. Chand. 1988. Validity of the environmental chloride method for recharge evaluation of coastal aquifers in India. J. of Hydrology. v. 99, pp. 349-366.

Sukhija, B. S., P. Nagabushanam, and D. V. Reddy. 1996. Ground water recharge in semiarid regions of India: An overview of results obtained using tracers. Hydrogeology Journal. v. 4, no. 3, pp. 50-71.

Taniguchi, M. and M. L. Sharma. 1990. Solute and heat transport experiments for estimating recharge rate. J. of Hydrology. v. 119, pp. 57-69.

B. S. Sukhija, D. V. Reddy, M.V. Nandakumar, and Rama, National Geophysical Research Institute, Uppal Road, Hyderabad, 500 007, India.

Received July 1995, revised February 1996, accepted March 1996.

Copyright Ground Water Publishing Company Jan/Feb 1997

Provided by ProQuest Information and Learning Company. All rights Reserved