ParametrizationTemporal
This class gives a parametrization of time depending parameters (gravity field, positions, ...). It will be used to set up the design matrix in a least squares adjustment. If multiple parametrizations are given the coefficients in the parameter vector are sequently appended.
Useally time intervals are defined half open meaning the last time belongs not to the interval. This behaviour can be changed for the last interval with includeLastTime.
Constant
Represents a parameter being constant in time in each interval.
The parameter names are *:*:*:<interval>
.
Name | Type | Annotation |
---|---|---|
interval | timeSeries | |
includeLastTime | boolean |
Trend
A time variable function is given by a linear trend \[ f(x,t) = \frac{1}{T}(t-t_0) \cdot f_t(x), \]with $t_0$ is timeStart and $T$ is timeStep in days. A constant term is not included and must added separately.
The parameter name is *:*:trend.<timeStep(days)>*(t-<timeStart>):*
.
Name | Type | Annotation |
---|---|---|
timeStart | time | reference time |
timeStep | time |
Splines
A time variable function is given by \[ f(x,t) = \sum_i f_i(x)\Psi_i(t), \]with the (spatial) coefficients $f_i(x)$ as parameters and the temporal basis functions $\Psi_i(t)$. Basis splines are defined as polynomials of degree $n$ in intervals between nodal points in time $t_i$, for details see basis splines.
The parameters are ordered timewise. First all parameters of $f_{i=1}(x)$ then $f_{i=2}(x)$ and so on. The total parameter count in each interval is $N=N_t+d-1$, where $N_t$ is the count of time points from timeSeries in each interval and $d$ is the degree.
The parameter names are *:*:spline.n<degree>:<interval of each spline>
.
Name | Type | Annotation |
---|---|---|
degree | uint | degree of splines |
timeSeries | timeSeries | nodal points in time domain |
intervals | timeSeries | |
includeLastTime | boolean |
Polynomial
A time variable function is represented by Legendre polynomials in each interval. The time is normed to $[-1,1)$ in each interval.
The total parameter count is $(N+1)M$,
where $N$ is the polynmial degree and $M$ the number of intervals with
the parameter names *:*:legendrePolynomial.n<degree>:<interval>
.
Name | Type | Annotation |
---|---|---|
polynomialDegree | uint | polynomial degree |
interval | timeSeries | intervals of polynomials |
includeLastTime | boolean |
Oscillation
A time variable function is given by a oscillation \[ f(x,t) = f^c(\M x)\cos(\omega_i(t)) + f^s(\M x)\sin(\omega_i(t)) \]with $\omega_i=\frac{2\pi}{T_i}(t-t_0)$, $t_0$ is timeStart and $T$ is timePeriod in days.
The parameter names are *:*:oscillation.cos(2*pi/<period(days)>*(t-<timeStart>)):*
and *:*:oscillation.sin(2*pi/<period(days)>*(t-<timeStart>)):*
.
Name | Type | Annotation |
---|---|---|
period | time | [day] |
time0 | time | reference time |
Fourier
A time variable function is given by a fourier expansion \[ f(x,t) = \sum_{m=1}^M f_m^c(\M x)\cos(2\pi m \tau) + f_m^s(\M x)\sin(2\pi m \tau) \]with the normalized time \[ \tau = \frac{t-t_A}{t_B-t_A}, \]and $t_A$ is timeStart, $t_B$ is timeEnd in each interval and $M$ is the fourierDegree.
The total parameter count is $2MN$, where $N$ is the number of intervals.
The parameters are sorted in following order: $f_1^c, f_1^s, f_2^c, \ldots$ with
the parameter names *:*:fourier.cos(<m>*x):<interval>
and *:*:fourier.sin(<m>*x):<interval>
.
Name | Type | Annotation |
---|---|---|
fourierDegree | uint | |
interval | timeSeries | intervals of fourier series |
includeLastTime | boolean |
DoodsonHarmonic
The time variable function is given by a fourier expansion \[ f(x,t) = \sum_{i} f_i^c(x)\cos(\Theta_i(t)) + f_i^s(x)\sin(\Theta_i(t)), \]where $\Theta_i(t)$ are the arguments of the tide constituents $i$ \[ \Theta_i(t) = \sum_{k=1}^6 n_i^k\beta_k(t), \]where $\beta_k(t)$ are the Doodson's fundamental arguments ($\tau,s,h,p,N',p_s$) and $n_i^k$ are the Doodson multipliers for the term at frequency $i$. The multipliers must be given by doodson coded as Doodson number (e.g. 255.555) or as names intoduced by Darwin (e.g. M2).
The major constituents given by doodson can be used to interpolate minor tidal constituents using the file inputfileAdmittance. This file can be created with DoodsonHarmonicsCalculateAdmittance.
The total parameter count is $2N$ with $N$ the number of doodson frequencies.
The parameters are sorted in following order: $f_1^c, f_1^s, f_2^c, \ldots$ with
the parameter names *:*:doodson.cos(<doodsonName>):*
and *:*:doodson.sin(<doodsonName>):*
.
Name | Type | Annotation |
---|---|---|
doodson | doodson | code number (e.g. 255.555) or darwin name (e.g. M2) |
inputfileAdmittance | filename | interpolation of minor constituents |