getRhoFirstOrderFun: Produces a function that uses a first-order autocorrelation...
In McDowellLab/loadflex: Models and Tools for Watershed Flux Estimates

Produces a function of the form described in correlations-1D.

getRhoFirstOrderFun accepts an assumed or empirically estimated value of ρ (rho), the correlation coefficient in the first-order autocorrelation model:

cor(y(t1), y(t2)) = ρ^τ

where

τ = y(t2) - y(t1)

and τ is in the same units as those used to estimate ρ (rho). To ensure that the units are the same, this function also accepts a time.step argument indicating the interval between two successive values in the time series used to estimate ρ (rho). The function returned by getRhoFirstOrderFun takes responsibility for matching the units of date1 and date2 to those of rho.

1	getRhoFirstOrderFun(rho, time.step = as.difftime(1, units = "hours"))

`rho`	The coefficient of the first-order autocorrelation model, either asserted by the user or estimated by `estimateRho`.
`time.step`	difftime. The time.step of the time series used to estimate `rho`. It is essential to use the right units; `as.difftime(1, units="hours")` is quite different from `as.difftime(1, units="days")`.

A function that accepts two date arguments and returns the predicted correlation between those two dates. One of the date arguments may be a vector as long as the other is scalar. This function takes the form, function(date1, date2) with arguments as defined for rhoEqualDates and rho1DayBand.

McDowellLab/loadflex documentation built on May 8, 2019, 9:48 a.m.