Source: UNISDR Disaster Risk Reduction
Nepal has developed an impact-based forecasting tool for floods, earthquakes, forest fires and air pollution, with historical loss data informing the impact estimates. Originally designed as a prototype for riverine floods and piloted at two river stations in West Nepal, the tool has since developed and improved and is designed for near-real-time estimates of impact, providing visualisations of hotspots for potential damage and loss
The National Disaster Risk Reduction and Management Authority of the Government of Nepal oversees and manages the national disaster information management system, the Building Information Against Disaster (BIPAD) portal. The system consists of separate platforms for national, provincial, and municipal governments, and features several modules, including risk information as well as damage and loss data.
Recently, a new module on impact-based forecasting has been developed and integrated into the portal (Figure 6).
The module quantifies and visualises three main elements in a dashboard: exposed buildings, road and land use in flood-prone regions using OpenStreetMap data overlayed with flood hazard maps; at-risk households, using household-level data collected by the Nepal Red Cross Society and scoring risk based on the INFORM Index; and potential flood impact. The potential flood impact is estimated including the use of historical disaster loss data as a reference.
The prototype module is being tested and planned improvements include the integration of local forecasts from the Department of Hydrology and Meteorology with higher accuracy for shorter lead times. For this, more reliable and locally available flood forecast data needs to be integrated into the portal along with flood extent maps.
Recent studies have shown that risk model outputs for early warning become more sensitive to vulnerability parameters as compared to hazard, the shorter the lead time. This means, while hazard-based risk forecasts provide useful information in the weeks of even several days ahead of an event, pre-existing vulnerability and historical impact data becomes more important for impact estimates one or two days or just hours ahead of the event.