sweq_simulate: Simulate from the modified SWEQ model
In robertsy/assimilr: Local Ensemble Kalman Particle Filters

Description Usage Arguments Value Examples

Simulate a time series of states and observations at a given frequency from the modified SWEQ model. The parameters and options follow the original implementation of the author in Python. Here the model is implemented in fortran90 for better performance.

sweq_simulate(duration, freq, ndim, umean = 0, unoise = FALSE, topo = 1,
  noiseswitch = 1, kh = 25000, ku = kh, kr = 200, noisefreq = 1,
  thres.rain = 0.005, sigr = 0.0025, sigu = 0.01, alpha = 1/4000,
  beta = 3, norain = TRUE, aircraft = 0, rgauss = FALSE,
  R.sigr = sigr, R.sigu = sigu, seed = sample.int(1000, 1), hc = 90.02,
  hr = 90.4, state0 = NA)

`duration`	the total integration time
`freq`	the frequency at which the process is recorded and observed
`ndim`	dimension of the 1-dim domain
`umean`	Initial wind field
`topo`	1=nothing, 2=one mountain, 3=5mountains
`kh, ku, kr`	diffusion parameters
`thres.rain`	threshold over which rain is observed
`sigr, sigu`	standard deviation of r and u observations
`alpha,`	beta model parameters
`R.sigr,`	R.sigu are used to create the R matrix (not necessarily equal to sigr and sigu)
`seed`	random seed (pass as NA if set externally)
`hc, hr`	critical heights for clouds and rain formation
`state0`	possible initial state at time 0
`unoise:`	TRUE/FALSE wind perturbation
`noiseswitch=1`	for adding random plumes
`noisefreq=1`	probability of random plume at each integration cycle
`norain:`	should no-rain observations be created

a list with everything needed for assimilation and plotting

ndim <- 300
freq <- 60
duration <- 12*freq
kh <- 25000; ku<- 25000; kr<- 200
alpha <- 1/4000; beta <- 3; hr <- 90.4; hc <- 90.02
train <- 0.005;
sigr <- 0.1
sigu <- 0.0025
R.sigu <- sigu
R.sigr <- 0.025
norain <- TRUE ## TRUE=assimilate norain observations
set.seed(1)
sweq_run <- sweq_simulate(duration, freq, ndim,
                          alpha=alpha, beta=beta,
                          noisefreq = 1, umean = 0,
                          hr=hr, hc=hc,
                          sigr=sigr, sigu=sigu,
                          R.sigr=R.sigr, R.sigu=R.sigu,
                          norain=norain,
                          thres.rain=train)
sweq_ggplot(sweq_run$state.ts[1,], obs=sweq_run$y.ts[[1]])