Mongabay · Natuur

Community Warnings Outperform Official System in Nepal Floods

During the August 26, 2026 Nepal floods, informal, community-based warnings proved more effective than the official SMS alert system, which was delayed, limited to Nepali, and lacked specific guidance. The floods resulted in 987 deaths and 3,916 missing persons.

Nepal's Department of Hydrology and Meteorology can issue forecasts but lacks the legal authority to mandate evacuations, leaving decisions to individuals. A 2022 study highlighted similar failures after a 2021 disaster, recommending binding Standard Operating Procedures that link forecasts to actions, a recommendation that remains unimplemented.

Existing solutions like cell broadcast technology, which can reach a zone in seconds in multiple languages, and scaled-up community-based networks offer potential improvements but have not been funded or implemented.

On August 26, head teacher Rajendra Dawadi received timely, urgent calls about an approaching flood, enabling him to evacuate his 900 students from Tribhuvan Trishuli Secondary School minutes before it was destroyed. This contrasted with the state's sensor and SMS system, which was less reliable and slower.

As of September 3, the floods had claimed 1,252 lives, with 4,216 missing and approximately 1.6 million people affected across five districts. Communication researcher Bhuwan KC noted that while some areas received information, it lacked urgency and scale.

The official flood warning system relies on threshold-based river sensors designed for gradual monsoon flooding. The Department of Hydrology and Meteorology learned of the flood after a call from a local official, and its first mass alert at 9:15-9:16 a.m. was reportedly received by some residents much later, with one resident receiving a 9:35 a.m. message at 9:54 a.m.

The SMS alert sent at 9:35 a.m. to Bidur read: “A large flood has entered the Bhotekoshi River in Rasuwa from the Tibet side. Until further notice, residents along the banks of the Trishuli River (up to Muglin) are requested to remain in a safe location with high alert.” This generic message lacked the specific, urgent, and actionable details that made community warnings effective.

Researchers like Chandani Bhandari of the University of Geneva pointed out that generic SMS warnings failed to convey individual vulnerability and lacked exact locations, severity, context, and actionable instructions. Furthermore, alerts were only in Nepali, excluding foreign visitors.

A 2026 review of Nepal's risk communication research found that DHM's SMS alerts repeatedly lacked actionable guidance. This issue is systemic, with warnings often arriving too late or being too generic to prompt immediate action.

A deeper structural problem is that the Department of Hydrology and Meteorology's mandate covers forecasting and dissemination, not ordering evacuations. The authority to relocate people rests with disaster management committees, but laws do not mandate real-time evacuation orders based on flood forecasts.

This gap between forecasting and action was also noted in a 2022 study of a 2021 disaster, which called for legally binding directives and SOPs to automatically trigger a chain of action following forecasts. This was not observed during the recent floods.

Community-based flood warnings, which rely on social connections and trusted individuals, have historically proven more effective in Nepal than formal messages. However, these approaches have been marginalized in favor of high-tech systems.

A trust deficit exists due to routine, low-stakes SMS alerts that often result in no action, leading people to treat them as background noise. Conversely, messages from friends and family are more trusted.

Potential fixes include cell broadcast technology, which can alert all phones in a zone within seconds and in multiple languages, and reinvestment in community-based networks where volunteers monitor water levels and relay warnings. While a GSMA/ITU study on cell broadcast technology was completed months before the flood, it had not been implemented.

The success of community-based warnings, like those that saved the school children, relies on trusted individuals relaying urgent information. This informal, redundant, and human-centered approach proved faster and more effective than the official, tech-reliant system during the August 26 floods.

AI-samenvatting op basis van de bron.