Probabilistic storm surge hazard using a steady-state surge model for the Pearl River Delta Region, China

Storm surges caused by tropical cyclones (TCs) are one of the costliest threats to coastal communities in southern China. Numerical surge models remain computationally challenging when used to simulate the large number of TC events required for probabilistic hazard assessments at regional scale. The present study demonstrates the applicability of a simple 1D steady-state storm surge representation for such regional scale hazard assessment. The surge setups from wind shear stress and barometric pressure difference are calculated with the meteorological forcing derived from parametric wind models and TC track information. Being computationally efficient, the surge model results do not require further empirical coefficients derived from correlation against observed data as compared to the previous statistical and semi-empirical surge estimations. Using the Pearl River Delta (PRD) region in China as a case study, the root-mean-square errors between the estimated and reported peak storm surges along the PRD coastline are 0.37 m and 0.45 m using two different TC best track inputs, respectively, covering 118 observed surge records from 39 historical TC events. Probabilistic surge hazard maps are further developed for the PRD coastline using the two TC best track datasets covering 1951-2018 as inputs. The mean surge heights along the coastline are in the range of 1.5-3.2 m and 2.0-3.5 m under 100-year and 200-year return periods, respectively. Areas in the west and near the estuary outlet are more prone to higher surge levels due to more frequent TCs affecting the areas historically. Differences in TC characteristics exist between the two best track datasets, which gives rise to localised difference in surge heights along the PRD coastline. The maximum differences in the 100-year and 200-year return period surge levels from the two best track datasets are 0.51 m and 0.64 m, respectively.