NEWS

---

mevr 1.3.4

Changes

• Added a delta-method for the calculation of confidence intervals of the SMEV.

mevr 1.3.3 (2026-10-13)

Changes

• Added some edge case tests to increase overall coverage.

mevr 1.3.2

Changes

• Force data.table > 1.18.0 to ensure proper handling of NA values in large datasets

mevr 1.3.1

Changes

• The data argument in fsmev now also accepts a vector.
• The set of independent storm events identified via event_separation now remains invariant across durations.

mevr 1.3.0

Changes

• New argument left_cens in fsmev() and fmev(). With left_cens = 0.9, c and w come from the top 10 % of events while n still counts all of them.
• Ordinary_events() no longer drops rows; missing values are handled inside the rolling sum.
• Bootstrap confidence intervals now refit with the object's own method, threshold and left_cens; before, they always refitted with default PWM.

mevr 1.2.3

Changes

• Included some tests for new functions for storm separation and ordinary events definition.

mevr 1.2.2

Changes

• Two functions added for storm separation and definition of ordinary events.
• Functions fmsev, fmev and ftmev can now be used for sub-daily rainfall durations.

mevr 1.1.2

Changes

• The Weibull tail test now returns the result to the fsmev function.
• Some tests were added.

mevr 1.1.1 (2024-06-30)

Changes

• A Weibull tail test based on left-censoring the input data can be performed

prior to fitting a SMEV. This can drastly improve fitting results

• Some internal code improvements.

mevr 1.0.1

Changes

• NA values are always removed. Functions fsmev, fmev and ftmev throw warnings

if the data contains any NA values

• Some internal code improvements.
• Added more tests.

mevr 1.0.0 (2024-06-15)

• First release, submitted to CRAN.