Changes in Dependence between Drivers of Compound Flooding around the
Contiguous United States Coastline
Abstract
Low-lying coastal zones are prone to flooding from multiple drivers such
as storm surge (oceanographic), excessive river discharge (fluvial),
and/or surface runoff (pluvial). The flooding impacts can be
exacerbated, depending on local characteristics, when flooding is
intensified by concurrent (or successive) occurrence of multiple drivers
known as ‘compound flooding’. Recently, compound flooding drivers are
becoming more frequent and intense leading to more adverse impacts. In
this study, we carry out a continental scale analysis for the CONUS
coastline at locations with sufficiently long overlapping records to
characterize the changes in dependence and co-occurrence between the
compound flooding drivers over time. We also investigate the changes in
dependence over time during tropical and extratropical seasons. Lastly,
we assess how the dependence structure varies with time. We use
observations (gauge records) for the analysis. Dependence between
different pairs is assessed using co-occurrence counts and statistical
measures for dependence (Kendall’s rank correlation coefficient, τ). The
dependence structures (particularly the tails of bivariate
distributions) are compared using Kullback–Leibler (KL) Divergence to
assess if there are significant changes in tails of bivariate
distributions over time. This analysis provides a comprehensive
characterization of changes in compound flooding potential around the
CONUS coastline. This will provide insights on where and how compound
flooding potential has changed over time to be incorporated in flood
risk assessments and planning.