Arsenic is highly toxic in its inorganic form. People are exposed to this poison by consuming contaminated drinking water or using contaminated water for preparing food or irrigation of food crops and tabac. Long-term exposure to arsenic can cause cancer and skin lesions. It has also been associated with cardiovascular disease, kidney failure and diabetes. In utero and early childhood exposure has been linked to negative impacts on cognitive development and increased deaths in young adults. Already in 2018, these facts were presented by the World Health Organization WHO, who stated that at least 140 million people in 50 countries have been drinking water containing arsenic at levels above the WHO provisional guideline value of 10 µg/l. The whole fact sheet published by WHO can be found here.

Using machine learning to develop a global risk model

Michael Berg and his team from Eawag’s Water Resources and Drinking Water Department have already carried out various studies to determine the extent of arsenic-contaminated groundwater aquifers: Firstly, the researchers found toxic arsenic concentrations in groundwater during field work in Cambodia, Vietnam, and the Amazon region. They then developed regional arsenic risk maps for China, Pakistan and South East Asia.
Now researchers led by Eawag geophysicist Joel Podgorski have developed a global risk model. This is based on a self-learning algorithm that is fed with the latest data on geology, soil properties, climate and around 200,000 specific arsenic concentration measurements. “The result is the most accurate and detailed risk map to date on the global scale of groundwater arsenic pollution,” explains Podgorski, lead author of the study. It has now been published in the renowned journal “Science” and was co-financed by the Swiss Agency for Cooperation and Development (SDC).
In addition to the geological parameters, figures on population density and statistics on groundwater use are also included in the risk model. The researchers were thus able to calculate that between 94 and 220 million people are potentially affected by arsenic contamination in drinking water.
While some hotspots have been known for a long time, such as parts of South and South East Asia, they also identified previously unknown areas. For example, the hazard map shows parts of Central Asia, the Sahel and other regions of Africa as potentially arsenic contaminated regions.
However, the models have too low a resolution to determine the arsenic contamination of individual groundwater wells. “These maps serve as a basis for determining where targeted arsenic tests should be conducted,” says Michael Berg.

Free Groundwater Assessment Platform facilitates local research work

With the financial support of the Swiss Agency for Cooperation and Development (SDC), Eawag, together with Ernst Basler + Partner and Hydrosolutions, has developed an online solution that is available free of charge to all interested parties: the Groundwater Assessment Platform GAP (www.gapmaps.org). This gives experts from all over the world the opportunity to visualize their own data with relatively little effort and to create their own risk maps of arsenic and fluoride contamination.
The original publication appeared in Science and can be found here.