Climate change causes risk change.
CRA uses nonstationary risk analysis to quantify the nature and extent of the change.

The norm in climatology is time-varying properties or nonstationarities.
Sometimes we have time-constant or stationary properties.

One should therefore be cautious using stationary analysis because
this assumption may be violated.
Because of this, CRA normally uses nonstationary analyses.
However, CRA is also able to perform stationary analyses,
if the assumption is met and such an analysis is required by the client.

## Analysis types

- Nonstationary: Inhomogeneous Poisson process estimation
- Nonstationary: Extreme value distributions with time-dependent parameters
- Nonstationary: Statistical hypothesis tests
- Stationary: Fitting extreme value distributions

## Estimation and hypothesis test methods

- Maximum likelihood
- Nonparametric kernel
- Bootstrap simulations

## Robustness evaluation

This answers how strongly the result depends on made assumptions.

- Sensitivity analyses
- Monte Carlo tests

## Detection of extremes

- Nonparametric robust regression
- Declustering
- Block extremes
- Peaks-over-threshold