Integrating resident digital sketch maps with expert knowledge to assess spatial knowledge of flood risk: A case study of participatory mapping in Newport Beach, California

• Demonstrated public participation geographic information systems (PPGIS) methods.
• Compared local and expert spatial knowledge of flood hazards in Southern California.
• Agreement varied by the scale of analysis and by personal and contextual factors.
• Education, income, and homeownership were associated with greater agreement.
• Findings inform spatial aspects of flood risk planning and communication.

Public participation geographic information systems (PPGIS) have been increasingly used to assess resident spatial knowledge of environmental hazards and to validate and supplement expert estimates of hazardous areas with local knowledge, but few studies have demonstrated methods for directly comparing local and expert knowledge of the spatial distribution of hazards. This study collected PPGIS digital sketch maps of flood-prone areas from 166 residents living adjacent to the Newport Bay Estuary in Southern California to examine variations in spatial knowledge of flood risk. First, we assessed agreement among participants and found that residents of areas with a higher percentage of homeowner, older, and higher income residents had greater agreement regarding areas at risk of flooding. Second, we introduced composite indices to assess the agreement between participant sketches of flood-prone areas with modeled estimates of the distribution of flood hazards, and found that the level of agreement between local and expert knowledge varied by the scale of analysis and by personal and contextual factors. Respondents with higher educational attainment, household income, and homeownership were associated with greater agreement between resident sketch maps and expert estimates of hazardous areas. Results inform spatial aspects of flood risk planning and communication by demonstrating how digital sketch maps can be used to identify potential shortcomings of expert hazard models, as well as hazardous areas where resident risk perception may be weak.