Skip to content ↓

Arsenic in Bangladesh drinking water from wells may be linked to crop irrigation

MIT graduate student Khandaker Nurul Ashfaque tests water chemistry in Bangladesh.
MIT graduate student Khandaker Nurul Ashfaque tests water chemistry in Bangladesh.
Photo / Charles F. Harvey
MIT's Charles Harvey (center) tastes molasses prior to its injection deep into a Bangladesh aquifer.
MIT's Charles Harvey (center) tastes molasses prior to its injection deep into a Bangladesh aquifer.
Photo / Volker Niedan

A ruthless killer in Bangladesh's drinking water is making millions of people sick and may be causing as many as 3,000 deaths each year. That killer - naturally occurring arsenic in the water drawn from family wells - appears to have been released through a process involving crop irrigation, at least in one part of the country.

At a research site in the southern part of Bangladesh, scientists calculated that irrigation pumping, which began in the last several decades, has dramatically altered groundwater flow through the aquifer, and that the resulting changes to the chemistry of the groundwater have the potential to either increase or decrease arsenic levels, in a paper published in the Nov. 22 issue of Science.

"Our data indicate that the arsenic was mobilized largely by degradation of dissolved organic carbon by microbes. Some of the organic carbon appears to have been drawn into the aquifer by irrigation pumping," said Professor Charles F. Harvey, assistant professor of civil and environmental engineering and lead author of the paper. "But the effects of irrigation are complex, probably lowering arsenic concentrations in other areas. Curtailing irrigation pumping is not a solution."

Harvey is also an author of another paper providing an epidemiological analysis of arsenic-induced illness throughout Bangladesh. He and his co-authors conclude that by replacing a selected 31 percent of the country's most tainted wells with deeper wells, arsenic poisoning and arsenic-induced cancers would be reduced by about 70 percent, assuming that arsenic concentrations in deep wells remain low. That paper will appear next year in the Water Resources Research journal published by the American Geophysical Union.

"If the arsenic levels remain low at greater depths, and I think they will, then deeper wells are a good solution. But this should be studied," said Harvey, who added that the best immediate solution is to filter the water.

Arsenic poisoning (usually characterized by sores on the chest or blackened knotty palms), as well as cancers of the skin, lung, liver, bladder and pancreas, have been linked to arsenic in drinking water. In 1998, the World Bank agreed to provide Bangladesh a $32.4 million credit to develop a method of controlling the arsenic. But today, most Bangladeshis continue to drink arsenic-laced water. The World Bank describes the problem as one of the world's primary environmental challenges. The World Health Organization refers to it as "the largest mass poisoning of a population in history."

The mass poisoning began, sadly enough, with a well-meaning attempt to provide clean drinking water for Bangladeshis, who suffered from cholera and other diseases caused by bacteria in water taken from surface reservoirs. To remedy that problem, the Bangladesh government, with the help of international aid organizations, drilled between 6 and 10 million wells at depths ranging from 50 to 300 feet to provide clean, safe water for individual households.

At about the same time, farmers in this largely rural country began irrigating land so that rice, the country's main food staple, could be grown during the six dry months of the year when monsoon flooding abates.

Cholera deaths dropped. But about 10 years into the use of the tube wells, villagers started displaying symptoms consistent with arsenic-related illnesses, and cancers of the skin and internal organs became common.


It is widely believed that the arsenic was transported on sediment into the Ganges River delta from the Himalayan mountains, which border Bangladesh on the north. What scientists don't know is how the arsenic is released into water in the aquifers in such great quantities. WHO's guideline for a maximum amount of arsenic in drinking water is 10 parts per billion (ppb). Many of the wells in Bangladesh measure more than 50 ppb. If the arsenic is in the aquifer only at certain levels, then digging even deeper wells might solve the problem.

The Harvey research team selected a site 30 km south of the nation's capital city, Dhaka, and 7 km north of the Ganges River. They installed research wells and collected sediment cores, determining that the sediment was typical of southern Bangladesh.

They found a fairly even distribution of solid arsenic in a variety of forms. To determine whether it was this arsenic that was mobilized in the aquifer and to figure out how that might have happened, the scientists injected test substances into their sampling wells and then took water samples. One such test substance was molasses, which is rich in organic carbon. They discovered that arsenic levels quickly rose as the organic carbon or its degradation products mobilized arsenic from the sediment.

To find the source of the carbon, Harvey's team used radiocarbon dating to determine the age of the carbon isotopes in the groundwater at various depths. They found that very young and very old carbon were mixed, and concluded that water had carried young organic carbon (probably derived from untreated waste on the surface) deep into the aquifer.

Using Bangladesh census data to estimate arsenic exposure distributions and epidemiological data from West Bengal and Taiwan - countries where the effects of arsenic poisoning from drinking water have been well-documented - the scientists predict that long-term exposure to present arsenic concentrations will result in approximately 1.2 million cases of hyperpigmentation, 600,000 cases of keratosis, 125,000 cases of skin cancer and 3,000 fatalities per year from internal cancers.

For the full story and more photos, click here.

A version of this article appeared in MIT Tech Talk on December 4, 2002.

Related Topics

More MIT News