A bicubic spline interpolation of unequally spaced data

✍ Scribed by M.A. Christie; K.J.M. Moriarty

• Publisher: Elsevier Science
• Year: 1979
• Tongue: English
• Weight: 780 KB
• Volume: 17
• Category: Article
• ISSN: 0010-4655
• DOI: 10.1016/0010-4655(79)90098-5

✦ Synopsis

A theorist may wish to interpolate data known to be a func-Catalogue number: ACZG tion of two variables. Often the data are not known on a regular grid, but are distributed irregularly. The "best approx-Program obtainable from: CPC Program Library, Queen's imation" when interpolating data which can be assumed to University of Belfast, Northern Ireland (see application form be error free is with the bicubic spline method 1,21. in this issue)

Method of solution Computer: CDC 6600;Installation: University of London

We use de Boor's method and one dimensional cubic spline Computer Centre interpolation to calculate the coefficients of the spline in the rectangle [x 1,xt+i] X [y1, v~+i 1. We can then obtain an

Operating System: CDC SCOPE interpolated value for the function and its first derivatives.

Programming Language used: FORTRAN IV Restrictions on the complexity of the problem

The number of x-data points must be less than 25 and the High speed storage required: 5.4 kwords number of y-data points must be less than 10. These values can be changed by the user. The point (x, y) at which an No. of bits in a word: 60 interpolated value for the function is required must always lie in the rectangle [xi,xn] X Lyi,yml. The program makes Overlay structure: none no restrictions on the spacing of the data points on the x and y axes.

No. of magnetic tapes required: none

Typical running time Other peripherals used: card reader, lineprinter

The test run output took about 4.4 s.