fitACS

Fits a parametric autocorrelation structure (ACS) to empirical ACF values
using Nelder-Mead optimisation with MSE criterion.

Makes univariate, multivariate, or random fields simulations precise and simple. Just select the desired time series or random fields’ properties and it will do the rest. CoSMoS is based on the framework described in Papalexiou (2018, <doi:10.1016/j.advwatres.2018.02.013>), extended for random fields in Papalexiou and Serinaldi (2020, <doi:10.1029/2019WR026331>), and further advanced in Papalexiou et al. (2021, <doi:10.1029/2020WR029466>) to allow fine-scale space-time simulation of storms (or even cyclone-mimicking fields).

Kevin Shook

CoSMoS

Complete Stochastic Modelling Solution

Simon Michael Papalexiou

Francesco Serinaldi

Filip Strnad

Yannis Markonis

fitACS function

<dl><dt>acf</dt>
<dd>numeric vector of autocorrelation function values from lag 0</dd>
<dt>ID</dt>
<dd>character; ACS identifier (e.g. <code>"weibull"</code>, <code>"paretoII"</code>)</dd>
<dt>start</dt>
<dd>numeric vector of starting parameter values; if <code>NULL</code>,
all parameters start at 1</dd>
<dt>lag</dt>
<dd>integer; number of lags to use; if <code>NULL</code>, lags up to the
first value \(\le 0.01\) are used (or all lags if none drops below 0.01)</dd></dl>

Arguments

Autocorrelation structure fitting — fitACS

<dl>

<dt>acf</dt>
<dd>numeric vector of autocorrelation function values from lag 0</dd>


<dt>ID</dt>
<dd>character; ACS identifier (e.g. <code>"weibull"</code>, <code>"paretoII"</code>)</dd>


<dt>start</dt>
<dd>numeric vector of starting parameter values; if <code>NULL</code>,
all parameters start at 1</dd>


<dt>lag</dt>
<dd>integer; number of lags to use; if <code>NULL</code>, lags up to the
first value \(\le 0.01\) are used (or all lags if none drops below 0.01)</dd>

</dl>

fitACS: Autocorrelation structure fitting

Description

Usage

Value

Arguments

See Also

Examples