This dataset comes with the paper by Brunner et al. (2020): 'How probable is widespread flooding in the United States?
It consists of four main components:
(1) stochastic continuous streamflow simulations for the 671 catchments in the CAMELS dataset by Addor et al. (2017) in the Conterminous United States (CONUS);
(2) peak-over-threshold flood events for the catchments in this dataset for different threshold derived using observed and n=100 stochastically simulated time series;
(3) an R-script with the essential code needed to 
	(a) extract flood events using the approach used by Brunner et al. (2020), 
	(b) compute and visualize regional hazard using a river-basin and local perspective as in Brunner et al. (2020).
(4) shapefiles for the 671 catchment outlets and the 18 river basins used for the regional hazard analysis in Brunner et al. (2020).

The stochastic simulations (1) are provided in the folders data_stochastic_simulations_part_1 and data_stochastic_simulations_part_2. The two folders should be merged in a folder data_stochastic_simulations after download when using the R-script.
The folder contains one .Rdata file for each of the 671 catchments in the CAMELS dataset.
Each data object is structured as follows:
YYYY: year
MM: month
DD: day
Qobs: observed streamflow data (daily, ft^3/s)
r1-r100: 100 stochastic simulation runs derived using the stochastic model PRSim.wave by Brunner and Gilleland (2020) (daily, ft3/s).

The extracted flood events (2) are provided in the folder extracted_events for the 671 stations and the different thresholds.
_flood_events_stoch_sim_thresh_. The catchment_id corresonds to the hru_id in the shapefile described below.
Each .RData object is structured as follows:
Date: date of peak flood occurrence
start: start date of floods
end: end data of flood
duration: flood duration (days)
volume: flood volume (Mio m^3)
magnitude: flood magnitude (ft^3/s)
set: stochastic simulation run flood has been extracted from (r1-r100)

The R-script regional_flood_hazard.R contains code for conducting the regional flood hazard analysis performed in Brunner et al. (2020). It requires downloading all the folders provided via this dataset and creating a results folder. 
The main path location needs to be adjusted and the stochastic simulations (parts 1 and 2) merged in one folder called data_stochastic_simulations.

The folder shapefiles contains three shapefiles:
671 catchments in the CAMELS dataset (locations of outlets): HCDN_nhru_final_671.shp
Outline of conterminuous United States: tl_2017_us_state.shp
18 large river basins: US_river_basins
For an example of how to read these shapefiles into R, please refer to the R-script described above.

References:
Addor, N., A. J. Newman, N. Mizukami, and M. P. Clark (2017), The CAMELS data set: Catchment attributes and meteorology for large-sample studies, Hydrol. Earth Syst. Sci., 21(10), 5293–5313, doi:10.5194/hess-21-5293-2017.
Brunner, M. I., and E. Gilleland (2020), Stochastic simulation of streamflow and spatial extremes: a continuous, wavelet-based approach, Hydrol. Earth Syst. Sci. Discuss., in press, doi:10.5194/hess-2019-658.
Brunner, M. I., Papalexiou, S., Clark, M. P., & Gilleland, E. (2020). How probable is widespread flooding in the UnitedStates?.Water Resources Research, 56,e2020WR028096. https://doi.org/10.1029/2020WR028096