This NetCDF file was created for the manuscript: Extreme Precipitation Return Levels for Multiple Durations on a Global Scale, 
Gründemann et al., 2021. 

This file contain modified MSWEP-V2.2 data (Beck et al., 2019).
The python package mevpy (https://github.com/EnricoZorzetto/mevpy) was used in order to estimate the parameters and extremes.

The file "GPEX.nc" contains global estimates of extreme precipitation using four extreme value methods (GEV, POT, MEV and Gumbel), the
observed extremes, underlying parameters and variables for Gründemann et al. (2021) for 8 durations ranging between 3 hours and 10 days. 
The spatial resolution is 0.1 arc degrees, and the provided coordinates are the middle of the grid-cells.

GPEX.nc has the following coordinates:
- lat	latitude		The latitude coordinate: 1480 values and ranges between 89.95 and -57.95 degrees north
- lon	longitude		The longitude coordinate: 3600 values ranges between -179.95 and 179.95 degrees east
- dur 	event duration 		The duration of the extreme precipitation event: 3, 6, 12, 24, 48, 72, 120 and 240 hours
- tr	return period		The return period of the extreme precipitation: 2, 5, 10, 20, 39, 50, 100, 200, 500, 1000 years
- year 	year			38 values (ranging between 0 and 37), equal to the number of years of the MSWEP dataset

GPEX.nc has the following variables:	
- gev_estimate			GEV estimate			Extreme precipitation return levels estimated using GEV (mm)
- pot_estimate			POT estimate			Extreme precipitation return levels estimated using POT (mm)
- mev_estimate			MEV estimate			Extreme precipitation return levels estimated using MEV (mm)
- gumbel_estimate		Gumbel estimate			Extreme precipitation return levels estimated using Gumbel (mm)
- observed_estimate		Observed estimate		Observed extreme precipitation return levels (mm)
- gev_location			GEV location parameter		Location parameter of the GEV distribution   
- gev_scale			GEV scale parameter		Scale parameter of the GEV distribution   
- gev_shape			GEV shape parameter		Shape parameter of the GEV distribution   
- gev_heaviness			GEV heaviness			Heaviness amplification factor (h) of the GEV distribution for T10, T100 and T1000
- pot_location			POT location parameter		Location parameter for a GEV distribution, estimated by fitting the GP distribution to peak-over-threshold precipitation events
- pot_scale			POT scale parameter		Scale parameter for a GEV distribution, estimated by fitting the GP distribution to peak-over-threshold precipitation events
- pot_shape			POT shape parameter		Shape parameter for a GEV distribution, estimated by fitting the GP distribution to peak-over-threshold precipitation events
- pot_heaviness			POT heaviness			Heaviness amplification factor (h) of POT for T10, T100 and T1000
- mev_scale			MEV scale parameter		Scale parameter of the MEV distribution
- mev_shape			MEV shape parameter		Shape parameter of the MEV distribution
- mev_no_events			MEV number of events		n parameter of the MEV distribution, number of events per hydrological year
- mev_mean_scale		MEV mean scale parameter	Mean of the scale parameter of the MEV distribution
- mev_mean_shape		MEV mean shape parameter	Mean of the shape parameter of the MEV distribution
- mev_mean_no_events		MEV mean number of events	Mean of the n parameter of the MEV distribution, mean number of events per hydrological year
- mev_heaviness			MEV heaviness			Heaviness amplification factor (h) of MEV for T10, T100 and T1000
- gumbel_location		Gumbel location parameter	Location parameter of the Gumbel distribution
- gumbel_scale 			Gumbel scale parameter		Scale parameter of the Gumbel distribution
- annual_maximum		annual maximum precipitation	Annual maximum precipitation for each hydrological year, from highest to lowest (mm)
- hydroyear			start of hydrological year	Number indicating the month in which the hydrological year starts
- running_parameter		running parameter		Running parameter (hours) of the declustering method by Marra et al. (2018)
- mask				land mask			Mask used for this study to indicate the land cells (1) and ocean cells (0)

The Generalized Extreme Value (GEV) distribution is fitted using the annual maxima of each hydrological year.
The Peak-Over-Threshold (POT) method is estimated by fitting the Generalized Pareto (GP) distribution to peak-over-threshold (POT) precipitation events of on average 3 events per year.
The Metastatistical Extreme Value (MEV)-Weibull distribution is estimated for each hydrological year.
The Gumbel distribution is estimated using the annual maxima of each hydrological year.
The observed extreme precipitation return levels are derived from the MSWEP dataset (highest return level is 39 years)