This function is used in cases where there are numerical problems with the estimation of the WRTDS model. This mostly happens during bootstrap estimation or when the data sets are very large. In order to reduce the collinearity in the explanatory variables, some random noise is added to the time and log discharge variables in the Sample data frame.

jitterSam(Sam, V = 0.2)

Arguments

Sam

data frame with at least columns DecYear and LogQ

V

a multiplier for the sd of the LogQ jitter. for example V = 0.02, means that the sd of the LnQ jitter is 0.02*sdLQ

Value

SamR a data frame structured like the Sam data frame but with the time and discharge variables modified by adding random jitter

Examples

eList <- Choptank_eList Sample_jitter <- jitterSam(eList$Sample)