The Ganga Water Machine
Ashok Khosla
T he
Ganga River Basin is one of the largest single bodies of fresh water
in the world. In addition to the dozens of mighty rivers that flow
into the Ganga, the sedimentary aquifers hold a vast sea of fresh
water under the ground, spreading out from the banks along the
length of these rivers. Within India, the Gangetic Basin covers more
than 800,000 square kilometers, with the larger tributaries alone
accounting for well over 10,000 kilometers of flowing water.
Conjunctive use of ground and surface water,
a technology well developed and tested elsewhere (particularly in
California, Ohio and many other parts of the US), could bring about
a major revolution in the management of our water resources in
Northern India. In addition to creating the basis of vast new
economic opportunities, including huge increases in food and energy
production, improved flood and drought control, river transport
year-round, better salt balance and port maintenance, it would
provide ecosystem services of the highest value, transforming the
lives and well-being of an entire nation.
In a seminal paper entitled "The Ganges Water
Machine", published nearly thirty years ago in the prestigious
American journal Science (9 May 1975, Volume 188, pp 611-616 –
reprints of which are available from the editor), Roger Revelle and
V. Lakshminarayana showed how a carefully designed but simple method
of creating underground storage of water along the rivers of the
Ganga Basin could lead to dramatically better use of the region’s
resources. This would produce massive increases in food production
(by as much as a factor of three, they estimated at the time), more
uniform flows of river waters throughout the year, lessening the
possibilities of both floods and droughts, and the consequent
improvements in downstream river quality. These, in turn, would lead
to huge new opportunities for improving both the economy and ecology
of the region.
Here, in italics, are some excerpts from the
original paper:
The fundamental problems of land and water
development in the Ganges Plain arise from the highly seasonal flow
of the river and its tributaries. Nearly 84% of the rainfall occurs
from June through September, and 80% of the annual river flow takes
place during the 4 months of July through October.
The dry-season flow of the Ganges is barely
sufficient for the needs of India and Bangladesh. If irrigation with
either groundwater of surface water continues to be developed along
the lines of the present programs, the dry-season flow will be
continually be reduced. …Because of the steep slopes of the
Himalayan foothills and the flatness of the Ganges Plain, surface
sites for storage are scarce, [and costly]. On the other hand, there
are great possibilities for underground storage, which should be
relatively inexpensive.
There are at least five ways in which a
portion of the monsoon flows coud be stored underground.
Infiltration into the water table in the monsoon season could be
increased by (i) water spreading in the Terai; (ii) constructing
bunds at right angles to the flow lines in uncultivated fields to
slow down run-off and increase infiltration; (iii) pumping out the
underground aquifers during the dry season in the neighbourhood of
nallahs which carry water during the monsoon; (iv) pumping out
groundwater during the dry season along certain tributaries of the
Ganges to provide space for groundwater storage [during the
monsoon]; and (v) increasing seepage from irrigation canals during
the monsoon season by extending the network of canals [,particularly
leaky ones], distributaries and water courses for kharif irrigation
and pumping this seepage water during the dry season.
The method used to find the depression (or
lowering) of the water table during low flow seasons is given below:
The drawdown due to pumping in an acquifer is
given by
 |
Equation (1) |
Where the drawdown s = h1 — h(m), h1 is the
initial saturated thickness of the aquifer (m), h is the height of
the water table during pumping (m), Q is discharge (m3/day), T is
the coefficient of transmissibility (m2 / day), S is the storage
coefficient (dimensionless), t is time since pumping started (days),
and r is the distance from the pumping well (m).
Equation 1 is the solution of the differential
equation
 |
Equation (2) |
subject to certain simple boundary conditions.
Since Eq.2 is linear, the super-position principle can be used to
find the solution when more than one well is being pumped in the
aquifer.
The advantages identified by the authors
included: increased water available for irrigation during the rabi
crop, reducing the huge seasonal variation in the flows of the
tributaries and the main river, greatly reduced losses by
evaporation, and all the consequent benefits for river transport,
waste dilution, downstream uses, etc.
To work properly, the Ganga Water Machine must
be designed properly. The authors provide the hydrological science
that underlies this technology, which shows that the primary
interventions required are the building of carefully spaced "well
fields" on either side of the nallahs, canals and tributaries, and
operating them so as to pump down the aquifers during the dry season
and thus raising the surface stream flows considerably. During the
monsoons, the five measures identified above to increase
infiltration will refill the aquifers.
Obviously, considerable energy is needed to
run the Ganga Water Machine, but work by Development Alternatives
has shown that only a small fraction of the irrigation water
generated would be sufficient to grow the biomass needed to power
the Machine. The revenues in terms of additional agricultural
production alone would be much larger than the costs of all the
energy required.
Table depicting possible river reaches for
underground storage of monsoon flows
River |
Length of well
fields (km) |
Monsoon flow (X 106
ha-m) |
Dry-season flow
(X 106 ha-m) |
Ratio of dry-season to stored
flow (%) |
Stored underground |
Remaining in river |
1) |
Ganga to Yamuna |
720 |
1.3 |
0.1 |
0.3 |
23 |
2) |
Ramganga |
560 |
1.0 |
0.1 |
0.2 |
20 |
3) |
Yamuna system |
1440 |
2.6 |
0.3 |
0.7 |
27 |
4) |
Gumti and Sai |
610 |
1.1 |
0.1 |
0.3 |
27 |
5) |
Gagra, Sarda and Rapti |
1280 |
2.3 |
3.2 |
1.2 |
52 |
6) |
Son |
640 |
1.2 |
1.0 |
0.5 |
42 |
7) |
Buhri Gandak and Baghmati |
720 |
1.3 |
3.2 |
1.0 |
77 |
8) |
Below the Kosi, including
Mahananda |
400 |
0.7 |
1.1 |
0.4 |
57 |
|
Total |
6370 |
11.5 |
9.1 |
4.6 |
42 |
|
Sum of 1 to 4 |
3330 |
6.0 |
0.6 |
1.5 |
25 |
|
Sum of 5 to 8 |
3040 |
5.5 |
8.5 |
3.1 |
56 |
The accompanying equations and graphs give a
glimpse into the mathematics underlying the working of the
technology, showing that within a period of 3 or 4 years, a properly
functioning Ganga Water Machine could establish a regime where the
the ground water table would have been drawn down sufficiently to
allow its conjunctive use with surface water to yield major
improvements in the year round availability of these resources. The
Table, taken from the original paper, gives an idea of the magnitude
of the underground storage and lean season flow improvements
possible for some of the tributaries of the Ganga – showing more
than 10 Million hectare metres storage – comparable to the amount
used for irrigation in the entire basin today.
The beauty of the Ganga Water Machine, at a
time when large-scale engineering works are so much in fashion, is
that although it is based on a grand vision with potential
region-wide impact, it can be implemented locally and piecemeal – an
excellent example of a "glocal" intervention, where we can "Think
Globally and Act Locally". This means that it allows for local
participation, control and access to benefits, gradual investment of
financial and engineering capital, low gestation periods and reduced
possibility for frictional losses such as political patronage and
corruption. And at the same time, it needs science and engineering
of the highest sophistication. q
Back to Contents
|