R-functions for Fitting the Metastatistical Extreme Value Distribution MEVD.
The MEVD assumes daily rainfall extremes being block maxima over a finite and stochastically variable number of “ordinary events” which are defined as samples from the underlying distribution (Marani & Ignaccolo, 2015, Zorzetto et al., 2016).
The functions in this package can be used to fit the MEVD, its simplified sibling SMEV (Schellander et al., 2019, Marra et al., 2019) and the explicitly non-stationary approach TMEV (Falkensteiner et al., 2023) to data series.
The package also includes functions for the preparation of rainfall timeseries to the analysis of sub-daily durations. In particular, a timeseries can be separated into independent events based on a defined dry spell duration. The separated series can the be analysed for maxima within the events, which are defined as ordinary events of the sub-daily duration.
The R-package mevr was written during the development of the TMEV (Falkensteiner et al., 2023). See also this GitHub repository which contains the original code.
The easiest way to get mevr is to install it from CRAN
install.packages("mevr")To install the development version from GitHub
# install.packages("pak")
pak::pak("haraldschellander/mevr")