generateMTSFast: Faster simulation of multiple time series with approximately...

In TycheLab/CoSMoS: Complete Stochastic Modelling Solution

Description

For more details see section 6 in Serinaldi and Kilsby (2018), and section 2.4 in Papalexiou and Serinaldi (2020)

Usage

generateMTSFast(
  n,
  spacepoints,
  margdist,
  margarg,
  p0,
  distbounds = c(-Inf, Inf),
  stcsid,
  stcsarg,
  scalefactor = 1,
  anisotropyarg = list(phi1 = 1, phi2 = 1, theta = 0)
)

Arguments

n	number of fields (time steps) to simulate
spacepoints	matrix (d x 2) of coordinates (e.g. longitude and latitude) of d spatial locations (e.g. d gauge stations)
margdist	target marginal distribution of the field
margarg	list of marginal distribution arguments
p0	probability zero
distbounds	distribution bounds (default set to c(-Inf, Inf))
stcsid	spatiotemporal correlation structure ID
stcsarg	list of spatiotemporal correlation structure arguments
scalefactor	factor specifying the distance between the centers of two pixels (default set to 1)
anisotropyarg	list of arguments establishing spatial anisotropy. phi1 and phi2 control the stretch in two orthogonal directions (e.g., longitude and latitude) while the angle theta controls a counterclockwise rotation (default set to list(phi1 = 1, phi2 = 1 , theta = 0) for isotropic fields)

Details

generateMTSFast provides a faster approach to multivariate simulation compared to generateMTS by exploiting circulant embedding fast Fourier transformation. However, this approach is feasible only for approximately separable target spatiotemporal correlation functions (see section 6 in Serinaldi and Kilsby (2018)). generateMTSFast comprises fitting and simulation in a single function. Here, we give indicative CPU times for some settings, referring to a Windows 10 Pro x64 laptop with Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz, 4-core, 8 logical processors, and 32GB RAM.
CPU time:
d = 2500, n = 1000: ~58s
d = 2500, n = 10000: ~160s
d = 10000, n = 1000: ~2955s (~50min)

Examples

coord <- cbind(runif(4)*30, runif(4)*30)

sim <- generateMTSFast(
    n = 50,
    spacepoints = coord,
    p0 = 0.7,
    margdist ='paretoII',
    margarg = list(scale = 1,
                   shape = .3),
    stcsarg = list(scfid = "weibull",
                   tcfid = "weibull",
                   scfarg = list(scale = 20,
                                 shape = 0.7),
                   tcfarg = list(scale = 1.1,
                                 shape = 0.8))
)

TycheLab/CoSMoS documentation built on Feb. 2, 2021, 3:49 p.m.