Mapping and Portraying Inundation Uncertainty of Bathtub-Type Models

Data errors are an unavoidable reality in maps of present conditions, but with the prevalence of ease-of-use formats and software these errors are becoming less evident. In earlier maps, the resolution or scale of the information and its presentation on a physical medium (e.g., contours) could inherently convey a level of vagueness that corresponded to the accuracy limitations. Maps of modeled output have additional accuracy considerations, especially if extrapolating future or potential events (e.g., 100-year storm). Sea-level rise (SLR) mapping falls into this category but is also highly dependent on present topographic conditions. SLR and other ecological models use continuous surfaces or digital surface models to generate derived information; however, presentation of the uncertainty information can be difficult or confusing. A technique that conveys uncertainty boundaries for a given confidence level was developed for the National Oceanic and Atmospheric Administration Coastal Services Center's SLR and inundation-visualization tool. The technique and definition of uncertainty levels described herein varies from other common methods, but the use of root-mean-square error (RMSE) to derive the assessment is similar. The technique uses the reported RMSE of both elevation and tidal surface and their relationship to a normal distribution. In this way, user-defined confidence levels can be used to map uncertainty both above and below the deterministic value produced in typical single-surface or bathtub models. The technique is used in this context for SLR and inundation mapping but also has applicability in mapping other phenomena.