Search for: All records

This paper proposes a flood risk visualization method that is (1) readily transferable (2) hyperlocal, (3) computationally inexpensive, and (4) geometrically accurate. This proposal is for risk communication, to provide high-resolution, three-dimensional flood visualization at the sub-meter level. The method couples a laser scanning point cloud with algorithms that produce textured floodwaters, achieved through compounding multiple sine functions in a graphics shader. This hyper-local approach to visualization is enhanced by the ability to portray changes in (i) watercolor, (ii) texture, and (iii) motion (including dynamic heights) for various flood prediction scenarios. Through decoupling physics-based predictions from the visualization, a dynamic, flood risk viewer was produced with modest processing resources involving only a single, quad-core processor with a frequency around 4.30 GHz and with no graphics card. The system offers several major advantages. (1) The approach enables its use on a browser or with inexpensive, virtual reality hardware and, thus, promotes local dissemination for flood risk communication, planning, and mitigation. (2) The approach can be used for any scenario where water interfaces with the built environment, including inside of pipes. (3) When tested for a coastal inundation scenario from a hurricane, 92% of the neighborhood participants found it to be more effective in communicating flood risk than traditional 2D mapping flood warnings provided by governmental authorities.