makePeakMatrix: Make Peak Lower- and Upper-Bounds Matrix from Peak Discharge...
In MGBT: Multiple Grubbs-Beck Low-Outlier Test

makePeakMatrix

R Documentation

Make Peak Lower- and Upper-Bounds Matrix from Peak Discharge Qualification Codes

Description

Make peak lower- and upper-bounds matrix from peak discharge qualification codes for left- and right-bounds (limits) based on the less than and greater than censoring.

Usage

makePeakMatrix(x, lo=0, hi=NA, eps=sqrt(.Machine$double.eps), wypad=FALSE)

Arguments

`x`	A `data.frame` having been run through `splitPeakCodes` first, but if columns resulting from those two functions are are not detected, then the water year column and code split are made internally and subset to `appearsSystematic` but not returned;
`lo`	The lower (left) bounds for `isCode4` peaks, but if it is `NA`, then the minimum of the real peak streamflows minus the `eps` is used with protection against going below zero;
`hi`	The upper (right) bounds for `isCode8` peaks, but if it is `NA`, then the maximum of the real peak streamflows plus the `eps` is used;
`eps`	A small number subtracted from the lower (left) or added to the upper (right) bounds when computed internally when `lo` and `hi`, respectively, are `NA`; and
`wypad`	A logical to left append two columns of `NA` to the lower (right) and upper (right) bounds. This feature is provided to support an input structure used within the copBasic package.

Value

A two- or four-column matrix is returned depending on argument wypad with row names set as the water year.

Author(s)

W.H. Asquith

Examples

## Not run: 
  # Example of a streamflow site with several less and greater than annual peaks.
  site <- "08136220"; # another good test is "02370750"
  r <- 1E3; par(ljoin=1); myblu <- rgb(0, 0.3, 0.7, 0.2)
  pk   <- splitPeakCodes(dataRetrieval::readNWISpeak(site, convert=FALSE))
  pk   <- pk[pk$appearsSystematic,]
  plotPeaks(pk, xlab="Water Year", ylab="Streamflow, cfs")
  mtext(site, line=0, font=2)
  zmat <- makePeakMatrix(pk, wypad=TRUE); Y <- pk$water_yr; Q <- pk$peak_va
  S1 <- copBasic::wolfCOPtest(Y,Q, zmat=zmat, rndphi=r)
  d1 <- S1$table # plot sigmas and p-values based on z-matrix (using codes 4 and 8)
  S2 <- copBasic::wolfCOPtest(Y,Q, zmat=NULL, rndphi=r, ties.method="average")
  d2 <- S2$table # plot sigmas and p-values of a potentially naive ties averaging.
  S3 <- copBasic::wolfCOPtest(Y,Q, zmat=NULL, rndphi=r, ties.method="random" )
  d3 <- S3$table # plot sigmas and p-values by randomization of the ties
  plot(  d1$sigmas, d1$sigmas_pvl, pch=16, col=myblu, xlab="Sigma", ylab="p-value",
         xlim=c(0,1), ylim=c(0,1))
  points(d2$sigmas, d2$sigmas_pvl, pch=25, col="green4", cex=2.1, bg="white", lwd=2)
  points(d3$sigmas, d3$sigmas_pvl, pch=24, col="red",    cex=1.4, bg="white", lwd=2)
  lines( d1$sigmas, d1$sigmas_pvl, col="blue")
  mtext(site, line=1, font=2) #
## End(Not run)

MGBT documentation built on Oct. 5, 2026, 9:07 a.m.