plotlag {MQMF}

plotlag plots the effect of a lag between two variables

Description

the use of the base function ccf can suggest a lagged relationship between a driver variable and a react(ing) variable. For example, cpue may respond to catches in a negative manner after a lag of a few years. One looks for a negative lag, which would imply that the driver variable influences the react(ing) variable after the given lag has passed. The lag is always assumed to be based on yearly intervals, though this can be changed.

Usage

plotlag(
  x,
  driver = "catch",
  react = "cpue",
  lag = 0,
  interval = "year",
  filename = "",
  resol = 200,
  fnt = 7
)

Arguments

x

the matrix containing columns of the named variables. It must contain columns with the same names as the driver and react(ing) variables

driver

the variable doing the influencing

react

the variable being influenced

lag

the time lag before the influence is felt

interval

the name of the time-interval variable, default='year'

filename

default is empty. If a filename is put here a .png file with that name will be put into the working directory.

resol

the resolution of the png file, defaults to 200 dpi

fnt

the font used in the plot and axes. Default=7, bold Times. Using 6 gives Times, 1 will give SansSerif, 2 = bold Sans

Value

a list containing some summary results, the anova of the linear model fitted in aov, and a summary of the linear model in summ

Examples

year <- 1985:2008
catch <- c(1018,742,868,715,585,532,566,611,548,499,479,428,657,481,645,
           961,940,912,955,935,940,952,1030,985)
cpue <- c(0.6008,0.6583,0.6791,0.6889,0.7134,0.7221,0.7602,0.7931,0.8582,
          0.8876,1.0126,1.1533,1.2326,1.2764,1.3307,1.3538,1.2648,1.2510,
          1.2069,1.1552,1.1238,1.1281,1.1113,1.0377)
dat <- cbind(year,catch,cpue)
out <- plotlag(dat,driver="catch",react="cpue",lag=7)
round(out$results,5)
out$summ