You are on page 1of 4

STATGRAPHICS – Rev.

7/3/2009

Bicubic Spline

Summary
The Bicubic Spline procedure connects a grid of (X,Y,Z) triplets with a smooth function
consisting of piecewise cubic polynomials. The function can then be used to determine an
interpolated value for Z at any specified X and Y within the range of the original data.

Beginning with a grid of m (X,Y,Z) triplets called “control points”, the procedure creates a
“natural” cubic spline, i.e., a function with continuous first and second derivatives across the
boundaries between the intervals on either side of the interior control points and second
derivatives equal to 0 at the endpoints. The coefficients of the spline are determined using a
tridiagonal algorithm.

Sample StatFolio: bspline.sgp

Sample Data:
The file bspline.sgd contains 3 columns (X, Y, and Z) for fitting a spline and 2 additional
columns (Xnew and Ynew) of values at which interpolation is desired.

X Y Z Xnew Ynew
1 10 8 3.5 25
1 20 3 4.5 35
1 30 7
1 40 5
2 10 7
2 20 2
2 30 3
2 40 4
3 10 6
3 20 7
3 30 5
3 40 3
4 10 3
4 20 8
4 30 12
4 40 2
5 10 5
5 20 10
5 30 20
5 40 2

Note that the values of X and Y define a rectangular grid.

 2009 by StatPoint Technologies, Inc. Bicubic Spline - 1


STATGRAPHICS – Rev. 7/3/2009

Data Input
The data required to create a bicubic spline consist of 3 numeric columns.

 Z: enter the Z values for each of m control points. The control points must be arranged over a
grid consisting of all combinations of selected values of X and Y. For example, if 5 values of
X are selected and 4 values of Y are selected, the grid would consist of m = 20 control points,
as illustrated in the sample data.
 X: enter the X values for each of m control points.
 Y: enter the Y values for each of m control points.
 Interpolate at X: enter the X value for n additional points with the range of X at which
interpolated values of Z are desired.
 Interpolate at Y: enter the Y value for n additional points with the range of Y at which
interpolated values of Z are desired.

 2009 by StatPoint Technologies, Inc. Bicubic Spline - 2


STATGRAPHICS – Rev. 7/3/2009

Analysis Summary
The Analysis Summary shows the number of input control points and a table of interpolated Z
values (if an entry was made in the Interpolate at X and Interpolate at Y fields on the data input
dialog box).

Bicubic Spline - Z versus X and Y


Z variable: Z
X variable: X
Y variable: Y
Number of data values = 20

Interpolated Values
Xnew Ynew Interpolation
3.5 25.0 8.70379
4.5 35.0 11.1282

Spline Plot
The Spline Plot shows the fitted spline and the m control points.

Spline Plot

20
16
12
Z

8
4 40
0 253035
0 20
1 2 3 4 1015 Y
5
X

Pane Options

 2009 by StatPoint Technologies, Inc. Bicubic Spline - 3


STATGRAPHICS – Rev. 7/3/2009
 Lines between X values – specify the number of additional lines to be added to the surface
between each unique value of X amongst the control points.
 Lines between Y values – specify the number of additional lines to be added to the surface
between each unique value of Y amongst the control points.
 Show points – check if you wish to display the control points in addition to the spline.
 Solid surface – check if you wish to display the spline as a solid surface rather than a wire
mesh.
 Resolution – specify the number of X locations at which the spline should be evaluated
when creating the plot. A larger resolution results in a smoother plot. You can set the default
resolution using the Preferences selection on the Edit menu.

Save Results
You can save the following results to the datasheet:

 Interpolated values: the interpolated values displayed on the Analysis Summary pane.

 2009 by StatPoint Technologies, Inc. Bicubic Spline - 4

You might also like