CoKriging

Command line

The CoKriging operation can be directly executed by typing one of the following expressions on the command line of the Main window:

  

OUTMAP

=

MapCoKriging(InputPointMapPred, InputPointMapCovar, Georef, SemiVarModelPred, SemiVarModelCovar, CrossVarModel, LimDist)

OUTMAP

=

MapCoKriging(InputPointMapPred, InputPointMapCovar, Georef, SemiVarModelPred, SemiVarModelCovar, CrossVarModel, LimDist, 1 | 0)

OUTMAP

=

MapCoKriging(InputPointMapPred, InputPointMapCovar, Georef, SemiVarModelPred, SemiVarModelCovar, CrossVarModel, LimDist, 1 | 0, min, max)

OUTMAP

=

MapCoKriging(InputPointMapPred, InputPointMapCovar, Georef, SemiVarModelPred, SemiVarModelCovar, CrossVarModel, LimDist, 1 | 0, min, max, No | Average | Firstval)

OUTMAP

=

MapCoKriging(InputPointMapPred, InputPointMapCovar, Georef, SemiVarModelPred, SemiVarModelCovar, CrossVarModel, LimDist, 1 | 0, min, max, Average | Firstval, Tolerance)

OUTMAP

=

MapCoKriging(InputPointMapPred, InputPointMapCovar, Georef, SemiVarModelPred, SemiVarModelCovar, CrossVarModel, LimDist, 1 | 0, min, max, Average | Firstval, Tolerance, plane | sphere)

where:

OUTMAP

is the name of the output raster map.

MapCoKriging

is the command to start the CoKriging operation.

InputPointMapPred

is the name of the input point map with the predictand (value domain) that is poorly sampled.
To use a value attribute column from an attribute table which is linked to an ID point map, use InputPointMap.Columnname.

InputPointMapCovar

is the name of the input point map with the covariable (value domain) that is well sampled.

Georef

is the name of an existing georeference for the output raster map.

SemiVarModelPred

Model(nugget, sill, range) | Power(nugget, slope, power)

This expression defines the semi-variogram model and its parameters that should be used for the predictand.

SemiVarModelCovar

Model(nugget, sill, range) | Power(nugget, slope, power)

This expression defines the semi-variogram model and its parameters that should be used for the covariable.

CrossVarModel

Model(nugget, sill, range) | Power(nugget, slope, power)

This expression defines the cross-variogram model and its parameters that should be used.

Model

Spherical | Exponential | Gaussian | Wave | RatQuad | Circular

nugget

value for the nugget, according to each semi- or cross-variogram (real value).

sill

value for the sill, according to each semi- or cross-variogram (real value).

range

value for the range, according to each semi- or cross-variogram (real value > 0).

slope

when using the Power model: value for the 'slope'. When pow is specified as 1 (i.e. thus using a linear model), then slope is the direction coefficient, i.e. Dg/Dh (real value >= 0).

pow

when using the Power model: an exponent 0 < real value < 2. When you use value 1, the Power model will become linear and the slope will be constant. If the power exponent is 2 the assumed stochastic model (‘randomness’) is not always justifiable and the interpolation can become pathological.

LimDist

a value for the limiting distance: points that are further away from an output pixel than the limiting distance will not be used in the Kriging equations. Normally, this is a value smaller than the range of the predictand semi-variogram model (0 <= real value < diagonal size of map).

1 | 0

Specify whether or not an error map has to be created. The error map will contain the square root of the Kriging error variance values, i.e. standard deviations per pixel. When this parameter is not specified, no error map will be created.

1

Create an error map. The error map will obtain the same name as specified for the output Kriging map followed by the additional string _Error.

0

Do not create an error map.

min, max

Optional parameters (integer values) to specify the minimum and maximum number of points that should be taken into account per Kriging estimate/prediction, i.e. the number of points within the limiting distance of an output pixel that should be taken into account in the calculation of the output value for that pixel.

  • When the minimum is not specified a minimum value 1 will be used. When, for an output pixel, less points than the specified minimum are found within the specified limiting distance, no Kriging is performed and the output pixel will become undefined. If the whole map contains less points than the specified minimum, an error message will appear and no output will be produced.
  • When the maximum is not specified, value 16 will be used. When, for an output pixel, more points than the specified maximum are found within the specified limiting distance, only the specified maximum number of points that are nearest to the output pixel will be used.

No | Average | Firstval

 

Choose how handle possible coinciding points or duplicates should be handled. When no method is specified, Average will be used.

No

No removal of duplicates. Mind: when there are coinciding points, the Kriging system may become unsolvable and will give undefined values for the pixels for which coinciding points are found.

Average

Values of duplicates are replaced by the average (arithmetic mean); the coordinates of the first point are taken as new position.

Firstval

Values of duplicates (and their coordinates) are replaced by those of the first point.

Tolerance

Optional parameter to specify the distance within which 2 sample points are considered to coincide. When not specified, a tolerance of 0.1 m will be used.

plane | sphere

optional parameter to calculate distances in a plane (plane), or on the sphere by using the projection of the coordinate system as used by the georeference of the output raster map (sphere). If not specified, distances will be calculated in a plane.

When the definition symbol = is used, a dependent output map is created; when the assignment symbol := is used, the dependency link is immediately broken after the output map has been calculated.

Example:

The following expression:

  

OutCoKr

=

MapCoKriging (MyPoints.MyAttrA, MyPoints.MyAttrB, MyGeoref, Spherical(10,70,800), Gaussian(30,80,1000), Power(20,40,2), 200, 1, 6, 20, Firstval, 0.2)

performs CoKriging:

The results of the expression are:

See also: