Typical initial model used in Data-Based Mechanistic modelling.

Typical initial model used in Data-Based Mechanistic modelling. Rainfall is scaled by corresponding streamflow values raised to a power. This SMA uses streamflow data, so can not be used for prediction.

dbm.sim(DATA, power, qlag = 0, scale = 1, return_state = FALSE)

Arguments

DATA

time-series-like object with columns P (precipitation) and Q (streamflow).

power

power to apply to streamflow values.

qlag

number of time steps to lag the streamflow (relative to rainfall) before multiplication.

scale

constant multiplier of the result, for mass balance. If this parameter is set to NA (as it is by default) in hydromad it will be set by mass balance calculation.

return_state

ignored.

Value

the simulated effective rainfall, a time series of the same length as the input series.

See also

hydromad(sma = "dbm") to work with models as objects (recommended).

Author

Felix Andrews felix@nfrac.org

Examples

## view default parameter ranges:
str(hydromad.options("dbm"))
#> List of 1
#>  $ dbm:List of 3
#>   ..$ power: num [1:2] 0 0.9
#>   ..$ qlag : num [1:2] -1 2
#>   ..$ scale: num NA

data(HydroTestData)
mod0 <- hydromad(HydroTestData, sma = "dbm", routing = "expuh")
mod0
#> 
#> Hydromad model with "dbm" SMA and "expuh" routing:
#> Start = 2000-01-01, End = 2000-03-31
#> 
#> SMA Parameters:
#>       lower upper  
#> power     0   0.9  
#> qlag     -1   2.0  
#> scale    NA    NA  
#> Routing Parameters:
#> NULL

## simulate with some arbitrary parameter values
mod1 <- update(mod0, power = 0.5, qlag = 0, tau_s = 10)

xyplot(cbind(HydroTestData, dbm.Q = predict(mod1)))


## show effect of increase/decrease in each parameter
parRanges <- list(power = c(0.01, 0.9), qlag = c(-1, 2))
parsims <- mapply(
  val = parRanges, nm = names(parRanges),
  FUN = function(val, nm) {
    lopar <- min(val)
    hipar <- max(val)
    names(lopar) <- names(hipar) <- nm
    fitted(runlist(
      decrease = update(mod1, newpars = lopar),
      increase = update(mod1, newpars = hipar)
    ))
  }, SIMPLIFY = FALSE
)

xyplot.list(parsims,
  superpose = TRUE, layout = c(1, NA),
  main = "Simple parameter perturbation example"
) +
  latticeExtra::layer(panel.lines(fitted(mod1), col = "grey", lwd = 2))