This folder contains all necessary scripts to run the dependence model as described in the paper "Assessing the spatial correlation of potential compound flooding in the United States" (https://doi.org/10.5194/egusphere-2025-2993). The structure is as follows: -> Folder 'Data_preprocessing' contains scripts for preprocessing GESLA3/NOAA total water levels and USGS river discharges --> preprocessing_gesla_waterlevel_detrending.py is used for detrending GESLA3 water levels to remove the effects of annual mean sea level variations; --> missing_water_levels.py is used for infilling missing water levels from simultaneous values at nearby stations; --> missing_discharge.py is used for downloading USGS river discharges and infilling missing values using 1) rating curves and 2) simultaneous values at nearby stations; --> decluster_SW.py is a function used for declustering time series using a storm window (SW) approach; --> combining_data_per_coast.py is used for combining data from individual locations to the defined study areas; --> GPD_thres.R is used for identifying the best threshold for a time series using the GPD_Threshold_Solari function of the 'MultiHazard' R package; --> identifying_historic_spatially_joint_events.py is used for identifying spatially joint events from observations and the script consists of the following steps: --->identifying events with compound flooding potential at individual locations --->matching the primary bivariate event with potential events at remaining locations -> Folder 'Dependence_model' contains scripts and functions for calculating the dependence structure and generating synthetic events --> function 'Migpd_Fit.R' is used for estimating componentwise semi-parametric marginal distributions where a GPD is fitted to peaks above a specified threshold and an empirical distribution is used for values below the threshold; --> functions 'mexTransform.R', 'revTransform.R', 'transFun.HT04.R', and 'u2gpd.R' are used for transforming marginal distributions onto a common scale; --> function 'predict.mex.conditioned.R' is used for generating random samples from the estimated dependence structure using a Monte-Carlo approach; --> 'Cal_dependence_automated_thres.R' is used for 1) estimating total number of events and the corresponding conditional variable, 2) calculating the dependence structure, and 3) generating synthetic events for each conditional variable. -> Folder 'Paper_figures' contains scripts for generating the figures presented in the manuscript and supplementary materials This model is developed at Department of Water & Climate Risk Institute for Environmental Studies Faculty of Science, Vrije Universiteit Amsterdam De Boelelaan 1111, 1081 HV Amsterdam Please send any questions or requests to huazhi.li@vu.nl