Skip to content

Built-in Variogram Models

Intro

\[ d \, - \, \text{distance} \]
\[ r \, - \, \text{range} \]
\[ c \, - \, \text{sill} \]
\[ V(d) \, - \, \text{variogram function} \]
\[ C(d) \, - \, \text{covariance function} \]

In general, the covariance function is defined as follows:

\[ C(d) = c - V(d) \]

For more information about variogram model parameters, please refer to this tutorial or see this section of the documentation.

Gaussian

\[ V(d) = c \cdot \left(1 - \exp\left(-\frac{d^2}{2 \cdot r^2}\right)\right) \]
\[ C(d) = c \cdot \exp\left(-\frac{d^2}{2 \cdot r^2}\right) \]

Exponential

\[ V(d) = c \cdot \left(1 - \exp\left(-\frac{|d|}{r}\right)\right) \]
\[ C(d) = c \cdot \exp\left(-\frac{|d|}{r}\right) \]

Spherical

\[ V(d) = \begin{cases} c \cdot \left(1.5 \frac{|d|}{r} - 0.5 \left(\frac{|d|}{r}\right)^3\right) & \text{for } 0 \leq |d| \leq r \\ c & \text{for } |d| > r \end{cases} \]
\[ C(d) = \begin{cases} c \cdot \left(1 - 1.5 \frac{|d|}{r} + 0.5 \left(\frac{|d|}{r}\right)^3\right) & \text{for } 0 \leq |d| \leq r \\ 0 & \text{for } |d| > r \end{cases} \]

Linear

\[ V(d) = \begin{cases} c \cdot \frac{|d|}{r} & \text{for } 0 \leq |d| \leq r \\ c & \text{for } |d| > r \end{cases} \]
\[ C(d) = \begin{cases} c \cdot \left(1 - \frac{|d|}{r}\right) & \text{for } 0 \leq |d| \leq r \\ 0 & \text{for } |d| > r \end{cases} \]