GOODVAR-Replace negative, zero and bad variance values

Usage:

goodvar in out bad neg zero

Description:

This routine checks the variance component of a datafile for values that are bad, negative, or zero and replaces them by values specified by the user. The specified value can be the null value ("!") which is translated into the bad value.

Parameters:

IN

The input datafile.

OUT

The output datafile.

BAD

The value which replaces bad values. Enter an exclamation mark to keep bad values. Bad values in VARIANCE or ERRORS are not allowed by Figaro. If DSA has to convert these arrays while mapping them, floating overflows or square roots of negative numbers may occur, and the application is liable to crash. [!]

NEG

The value which replaces negative values. Enter an exclamation mark to replace negative values with the bad value. Negative errors or variances are nonsense. Negative variances often will cause an application to crash because it takes the square root to calculate the error. [!]

ZERO

The value which replaces zeroes. Enter an exclamation mark to replace zeroes with the bad value. Errors of zero sometimes are reasonable or necessary for error propagation. In other instances they cause problems, because statistical weights often are the reciprocal of the variance. [!]

Authors:

HME: Horst Meyerdierks (UoE, Starlink)