The River Itself Warns Us: Local Knowledge of Flood Forecasting in the Gandaki River Basin, West Champaran, India

This report presents information on local knowledge by documenting various, sophisticated methods deployed by communities to forecast floods and heavy rainfall in the Gandaki River basin. It makes a point that strengthening such local knowledge systems will address the challenges of centralized and mechanized early flood warning systems and ensure their recognition and continuation.

This research also seeks to clarify its definition of local knowledge in an effort to move away from the interchangeable use of traditional/indigenous/local knowledge, a flexibility that has created ontological problems.

By documenting the complex interaction between local-knowledge based forecasting and official early warning systems, the study recognizes the following points:

• A high degree of gender blindness embedded in official early flood warning systems;
• Knowledge innovation practised by local communities that convert obtuse flood- related information to practical use through a complex process of triangulation;
• Change to local knowledge asymmetries being brought in through the rapid proliferation of mobile phones.

Finally, it advocates knowledge co-production between local and scientific communities, which would only serve to strengthen one another.