Newspapers as a validation proxy for GIS modeling in Fujairah, United Arab Emirates: identifying flood-prone areas.

The UN Office for Disaster Risk Reduction listed 10 reasons businesses should reduce their disaster exposure, including risk factoring, which cannot be achieved without historical data about hazards, their locations, magnitudes, and frequencies. Substantial hazard data are reported by newspapers, which could add value to disaster management decision making. In this study, a text-mining program extracted keywords related to floods' geographic location, date, and damages from newspaper analyses of flash floods in Fujairah, UAE, from 2000–2018. The paper describes extracting such information as well as geocoding and validating flood-prone areas generated through geographic information system (GIS) modeling. The generation of flood-prone areas was based on elevation, slope, land use, soil, and geology coupled with topographic wetness index, topographic position index, and curve number. Analytical Hierarchy Process (AHP) produced relative weight for each factor, and GIS map algebra generated flood-prone areas. AHP inclusion helped minimize weight subjectivity among various experts. Of all areas, 85% are considered medium and low flood-prone zones, mainly mountainous areas. However, the 15% that are high/very high are dominated by urban areas in low coastal plains, predisposing them to flash floods. Eighty-four percent of flood events reported by newspapers were in areas rated as high/very high flood-prone zones. In the absence of flood records, newspapers reports can be used as a reference. Policymakers should assess whether flood-prone area models offer accurate analyses. These findings are useful for organizations related to disaster management, urban planning, insurance, archiving, and documentation. [ABSTRACT FROM AUTHOR]