21-st century probabilistic daily storm surge projections for Japan's coasts

This unique data product provides daily storm surge projections for various Japanese coasts from 2015 to 2100 in a probabilistic manner (i.e., five distinct percentiles per day). The projections are derived using statistical downscaling with Bayesian Quantile Regression (BQR), incorporating atmospheric covariates from ten advanced Global Climate Models with 25 km resolution. Additionally, data for a historical baseline period (1981-2014) are also included for reference. The regression models were previously developed using ERA5 re-analysis data and high-quality tide gauge observations. These storm surge data can be useful to both practitioners and researchers interested in future coastal flood risk assessments for Japan. The R/Stan codes to obtain BQR parameters for an example tide gauge station are also given.