This research is a case study of the city of Chennai (formerly Madras), located in the state of Tamil Nadu, in South India. As per the 2001 census, 4.3 million people lived within municipal corporation boundaries; 6.4 million lived in the Chennai urban agglomeration (includes peri-urban towns, suburbs, and villages). Chennai is a particularly water-scarce city. In fact, it has the lowest water availability per capita of any large metropolitan area in India.
A public water utility, the Chennai Metropolitan Water Supply and Sewerage Board (called "Metrowater"), serves the municipal corporation area via a piped network. Almost all households in Chennai have some sort of access to public supply: private piped connections, yard handpumps or taps, public standpipes, or utility run "mobile supply" tankers. Outside city limits, peri-urban towns and villages are served by a patchwork of town and village supply schemes, and are mostly dependent on groundwater.
Metrowater obtains most of its water for city supply via three interconnected rain-fed reservoirs, along with well-fields located to the north of the city. In addition, Metrowater also gets water from two inter-basin projects: the inter-state Telugu Ganga Project (water is delivered into the city's reservoir system), and the newly commissioned intra-state Veeranam Project (water is delivered directly to pumping stations). The locations of the sources and the quanities supplied in a good year (by our estimates) are shown in the figure below.
Figure I. Sources of Supply to Chennai City, Including Utility and Non-utility Sources
Between 2002 and 2006, the quantity of water available from all these sources varied significantly from month to month. The quantity available to households ranged from 60 to 100 LPCD5 after pipeline losses, commercial needs were accounted for, and self-supply and supply from private tankers was factored in.
In 2003-2004, Chennai’s reservoirs went completely dry; the piped supply system was virtually shut down for almost a year. The entire city was supplied by “mobile supply”, utility run tankers that went from neighborhood to neighborhood delivering a lifeline supply of water (about 20 liters per capita per day). The cessation of piped supply for almost a year in a large Metropolitan area represented a crisis of severe magnitude and prompted speculation that the city might have to be evacuated if no water was made available soon.
A household survey conducted during the drought showed that over two thirds of households reported depending largely on their private wells. Over 6 percent purchased water from private tanker suppliers who trucked in water extracted from peri-urban agricultural wells.
As Chennai was suffering a severe water crisis, three very different solutions emerged to address Chennai’s water problems: the utility favored augmenting supply by building desalination plants, economists at the development banks were promoting efficiency improvement by raising tariffs and fixing leaky pipes, while environmental NGOs were promoting harvesting rainwater to recharge the aquifer. The problem was that no framework existed to compare the costs and benefits of this wide range of solutions.
The goal of this research is to figure out which of these solutions (or combination of solutions) would be best (efficient, equitable, sustainable) in solving Chennai’s water problems. We address these research goals by asking the following research questions
Explain currently observed trends:
To address the research goals laid out, we developed an integrated dynamic simulation model as shown in Figure II. The integrated model is a transient, spatially explicit model of the Chennai basin, which incorporates the linkage of three scales: user-scale consumer behavior, utility-scale operations, and basin-scale water resources simulation modeling.
The model was run for two periods: The historical run from 2002 to 2006 used to calibrate the model and develop insights on the Chennai water system, and forecasting runs from 2007 to 2025 were used to develop scenarios of Chennai's water supply situation in 2025 and test the effects of various policies.
Figure II. A conceptual model of the integrated simulation model.
The integrated model consists of five inter-linked modules. Each module is divided into sub-modules consisting of one or more linked equations, which are calibrated independently.
Spatial and Temporal Units
The model was run with a time-step of 3-months to capture seasonality. The spatial discretisation of the modules varied. The groundwater model was a grid-based MODFLOW model covering the Chennai-basin area of 50.8 km * 50.8 km. Each grid cell was 0.22 km * 0.22 km. The consumer and utility modules used census units (corporation zone within the city, census block outside).
The integrated model was run over the period from January 2002-January 2006. This period included a multi-year drought (2003-2004), as well as a year in which Chennai received the highest rainfall in recorded history (2005) as shown in Figure III.
Figure III. Annual Rainfall from 2002 through 2006.
When presenting model results, we offer comparisons using the two climatic extremes as reference periods: Jan-Apr 2004 (dry) and Jan-Apr 2006 (wet) following the record rains.
The historical run showed that the Chennai reservoir system is capacity-constrained (given inflows and diversions). This made utility supply highly variable and intermittent. In periods when utility supply was reduced due to lack of availability, consumers depended on private and community wells. As extractions increased and recharge was lower, groundwater levels fell. As the aquifer dried up, so did consumer wells, forcing customers to purchase tanker water. (Tanker water is an order of magnitude more expensive than the cost of groundwater or utility supply.) Thus, consumers suffered significant losses in well-being, which the model allowed us to quantify.
Chennai 2025 Baseline Scenarios
The model was then extrapolated to 2025 by extrapolating population, income, and land use changes. A 100 million liters per day desalination plant was assumed to be commissioned in 2009. The model used different rainfall scenarios by repeating stretches of the historical rainfall record in the period from 2008-2025.
The baseline scenarios yielded two interesting results. First, contrary to our expectations, increased urbanization and the consequent displacement of irrigated agriculture did not "free up" additional water for urban uses elsewhere. Instead, rising populations, incomes, and commercial and industrial growth consumed most of the water previously used by irrigated agriculture. Secondly, the 100 million liters per day desalination plant, assumed to be commissioned in 2009, did not alleviate a multi-year drought significantly. The utility supply still would have needed to be shut down, and the aquifer dried up to generate a big tanker market. However, if the desalination plant did prevent the situation from getting worse, it would allow the water utility to keep pace with the increase in demand due to rising populations and incomes.
Chennai 2025 Policy Scenarios
The model was then extrapolated to 2025 by extrapolating population, income, and land use changes. The model used different rainfall scenarios by repeating stretches of the historical rainfall record in the period from 2008-2025.
We compared and presented results from three different policy situations:
Key Insights from the Model
The model replicated all aspects of the 2003-2004 drought; the city’s reservoirs dried up and the piped supply system shut-down. As consumers became increasing dependent on private wells, the aquifer dried up and consumers were forced to purchase expensive private tanker water, causing much distress to consumers. The research predicts that given Chennai’s rainfall patterns and limited reservoir capacity, the severe water crisis which occurred in 2003-2004 is likely to occur again, even if the proposed 100 million liters per day desalination plant is constructed.
Building additional desalination plants would make more water available but is not cost-effective at current technology and energy prices. Instead, the model found that raising tariffs, fixing the leaky distribution network, and implementing aggressive rainwater harvesting are cost-effective and could significantly alleviate a future water crisis.
This research suggests adopting a “dual-quality” solution for water-scarce cities like Chennai. The dual-quality solution would require altering the plumbing code to encourage builders to keep water quality separate in buildings and houses. The utility will provide high-quality, high-cost, metered, 24*7 water supply via the piped distribution network for kitchen use (drinking, cooking, dish-washing), while consumers would continue to rely on private wells for bathroom use (flushing, bathing, clothes washing) of water.
Srinivasan, V., S. M. Gorelick, and L. Goulder. 2010. A hydrologic-economic modeling approach for analysis of urban water supply dynamics in Chennai, India, Water Resources Research. vol. 46, W07540, doi:10.1029/2009WR008693.
Srinivasan, V., S. M. Gorelick, and L. Goulder. 2010. Sustainable urban water supply in South India: Desalination, Efficiency Improvement, or Rainwater Harvesting?, Water Resources Research. vol. 46, doi:10.1029/2009WR008698.
Srinivasan, V., L. Goulder, and S.M. Gorelick. 2010. Factors determining informal tanker water markets in Chennai, India, Water International. vol. 35, no. 3, p. 254-269.
Srinivasan, V., K. Seto, R. Emerson, and S.M. Gorelick. 2013. The impact of urbanization on water vulnerability—A coupled human - environment system approach for Chennai, India, Global Environmental Change - Human and Policy Dimensions, vol. 23 , Issue: 1, 229-23 doi:10.1016/j.gloenvcha.2012.10.002