Probabilistic Mapping of Storm-induced Coastal Inundation for Climate Change Adaptation

Global warming is posing an imminent threat to coastal communities worldwide. Under the IPCC RCP8.5 scenario, we utilize hurricane events downscaled from a CMIP5 global climate model using the stochastic-deterministic method of Emanuel (2013, Proc. Nat. Acad. Sci.) in a pilot study to develop an inundation map with projected sea-level rise for the urban Honolulu coast. The downscaling is performed for a 20-year period from 2081 to 2100 to capture the ENSO, which strongly influences the hurricane activity in the Pacific. A total of 50 simulations provide a quasi-stationary dataset of 1000 years for probabilistic analysis of the flood hazards toward the end of the century. We utilize the meta-model Hakou, which is based on precomputed hurricane scenarios using ADCIRC, SWAN, and a 1D Boussinesq model (Kennedy et al., 2012, Ocean Modelling), to estimate the annual maximum inundation along the project coastline at the present sea level. Screening of the preliminary results identifies the most severe three events for detailed inundation modeling using the package of Li et al. (2014, Ocean Modelling) at the projected sea level. For each event, the third generation spectral model WAVEWATCH III of Tolman (2008, Ocean Modelling) provides the hurricane waves and the circulation model NEOWAVE of Yamazaki et al. (2009, 2011, Int. J. Num. Meth. Fluids) computes the surge using a system of telescopic nested grids from the open ocean to the project coastline. The output defines the boundary conditions and initial still-water elevation for computation of phase-resolving surf-zone and inundation processes using the 2D Boussinesq model of Roeber and Cheung (2012, Coastal Engineering). Each computed inundation event corresponds to an annual maximum, and with 1000 years of data, has an occurrence probability of 0.1% in a given year. Barring the tail of the distribution, aggregation of the three computed events allow delineation of the inundation zone with annual exceedance probability equal to or greater than 0.2% (equivalent to a 500-year return period). An immediate application is to assess the inventory of buildings and structures in Honolulu that would be exposed to increased flood risks due to climate change and identify potential revisions to the building code as part of the adaptation process.