Skip to contents

Conditionally sampling a two dimensional dataset

Con_Sampling_2D_Lag.Rd

Creates a data frame where the declustered excesses of a (conditioning) variable are paired with the maximum value of a second variable over a specified time-lag.

Usage

Con_Sampling_2D_Lag(
  Data_Detrend,
  Data_Declust,
  Con_Variable,
  u = 0.97,
  Thres = NA,
  Lag_Backward = 3,
  Lag_Forward = 3
)

Arguments

Data_Detrend

Data frame containing two at least partially concurrent time series, detrended if necessary. Time steps must be equally spaced, with missing values assigned NA. First object may be a "Date" object. Can be Dataframe_Combine output.

Data_Declust

Data frame containing two (independently) declustered at least partially concurrent time series. Time steps must be equally spaced, with missing values assigned NA. Columns must be in the same order as in Data_Detrend. First object may be a "Date" object. Can be Dataframe_Combine output.

Con_Variable

Column number (1 or 2) or the column name of the conditioning variable. Default is 1.

u

Threshold, as a quantile of the observations of the conditioning variable. Default is 0.97.

Thres

Threshold expressed on the original scale of the observations. Only one of u and Thres should be supplied. Default is NA.

Lag_Backward

Positive lag applied to variable not assigned as the Con_Variable. Default is 3

Lag_Forward

Negative lag to variable not assigned as the Con_Variable. Default is 3

Value

List comprising the specified Threshold as the quantile of the conditioning variable above which declustered excesses are paired with co-occurences of the other variable, the resulting two-dimensional sample data and Con_Variable the name of the conditioning variable. The index of the input dataset that correspond to the events of the conditioning variable x.con and the non-conditioning variable x.noncon in the conditonal sample are also provided.

Examples

S20.Rainfall<-Con_Sampling_2D_Lag(Data_Detrend=S20.Detrend.df[,-c(1,4)],
                                  Data_Declust=S20.Detrend.Declustered.df[,-c(1,4)],
                                  Con_Variable="Rainfall",u=0.97)