Researchers from six organizations have developed an early warning and artificial intelligence tool to predict water-related conflicts with the aim of preventing them and reacting preventively and effectively. It is developed in collaboration with the Netherlands Ministry of Foreign Affairs and Deltares, IHE Delft, International Alert, Centre for Strategic Studies in The Hague, Wetlands International and the World Resources Institute.
It is the first tool which uses environmental, rainfall and flood data combined with socio-economic variables. Specifically, NASA and European Space Agency satellites carry out the monitoring while the rest of the information comes from databases of governments, international agencies and research organizations.
The machine learning model is trained to identify patterns using historical data on violent conflict and political, social, economic, demographic and water risks. According to Charles Iceland of the World Resources Institute, “machine learning is able to detect patterns in data where humans cannot”
This tool has already predicted conflicts that are likely to occur by 2020 in hot spots in Africa, the Middle East and Southeast Asia. The developers say the success rate in identifying conflict zones where at least 10 deaths could occur is 86%. The growing global demand for water is already creating tensions between farming and urban communities and between population and governments.
Tensions are expected to increase as water scarcity becomes a reality for more people. According to the United Nations, up to 5 billion people could face water shortages by 2050.
1: S. Kamali Dehghan, The Guardia, ´Water wars: early warning tool uses climate data to predict conflicts hotspots´, https://www.theguardian.com/global-development/2020/jan/08/water-wars-early-warning-tool-uses-climate-data-to-predict-conflict-hotspots, (accessed 14 January 2020)