Compute return levels with confidence intervals — compute_return

Takes kriging predictions (NOAA GEV parameters and their covariance) and computes r-year return levels with 95% confidence intervals via the delta method and/or simulation from the predictive distribution.

Usage

compute_return_levels(
  predictions,
  r = 100,
  alpha = 0.05,
  method = "both",
  n_sim = 2500,
  seed = NULL
)

Arguments

predictions: An evfuse_predictions object from predict_krig.
r: Return period in years (default 100).
alpha: Confidence level for CIs (default 0.05 for 95% CIs).
method: One of "delta", "simulation", or "both" (default "both").
n_sim: Number of simulations for the simulation method (default 2500).
seed: Random seed for simulation method.

Value

A data frame with columns:

lon, lat: Coordinates of prediction sites.
return_level: Point estimate of the r-year return level.
se_delta: Standard error from the delta method (if computed).
ci_lower_delta, ci_upper_delta: Delta method CI bounds.
se_sim: Standard error from simulation (if computed).
ci_lower_sim, ci_upper_sim: Simulation-based CI bounds.