The 2014 Karnali River Floods in Western Nepal: Making Community Based Early Warning Systems Work When Data Is Lacking

Abstract:

Implementing Community Based Flood Early Warning System (EWS) in developing countries like Nepal is challenging. Complex topography and geology combined with a sparse network of river and rainfall gauges and little predictive meteorological capacity both nationally and regionally dramatically constrain EWS options. This paper provides a synopsis of the hydrological and meteorological conditions that led to flooding in the Karnali River, West Nepal during mid-August 2014, and analyses the effectiveness of flood EWS in the region.

On August 14-15, 2014, a large, slow moving weather system deposited record breaking rainfall in the foothills of the Karnali River catchment. Precipitation depths of 200 to 500 mm were recorded over a 24-hour period, which led to rapid rise of river heights. At the Chisapani river gauge station used for the existing EWS, where the Karnali River exits the Himalaya onto the Indo-Gangetic Plain, water levels rapidly exceeded the 11 meter danger level. Between 3 to 6 am, water levels rose from 11 to 16. 1 meters, well beyond the design height of 15 meters. Analysis suggests that 2014 floods may have been a one-in-1000 year event.

Starting with the onset of intense rainfall, the Chisapani gauge reader was in regular communication with downstream stakeholders and communities providing them with timely information regarding rising water level. This provided people just enough time to move to safe places with their livestock and key assets. Though households still lost substantial assets, without the EWS, floodwaters would have caught communities completely unaware and damage would almost certainly have been much worse. In particular, despite the complications associated with access to the Chisapani gauge and failure of critical communication nodes during the floods, EWS was instrumental in saving lives.

This study explores both the details of the flood event and performance of the early warning system, and identifies lessons learned to help strengthen flood response in Nepal other regions facing similar data and technology constraints. This research also underscores the need to move from observation to forecast based EWS, ideally coupled with hydrological and meteorological models that would provide flood estimates well in advance to vulnerable communities.