ASTM notes: Additional information concerning the ASTM procedure

Note of November 14, 2003:
This is a set of notes prepared by H.R. Olesen of NERI, Denmark. It is meant only for readers who wish to dive deeply into working with the ASTM procedure. Much of the information found here is not any longer relevant, as the software related to the methodology has changed since I wrote the notes.

These notes were originally - in 1998 - written as a service to readers who intended to work with the draft ASTM procedure for statistical evaluation of models.

The notes concern the experiences I have had while working with John Irwin’s programmes ASTM.FOR and NEWFIT.FOR in conjunction with Kincaid data. The notes were first prepared in spring 1998 while working with earlier versions of ASTM and NEWFIT. They have been revised in March 1999 to reflect that an updated version of NEWFIT (of February 12, 1999) has been released.

The information in the notes may still have some relevance, but if you want to work with the ASTM methodology, you should first contact John Irwin in order to obtain the most recent version of his data and tools.

Data and programmes are available "as is" (presently the system is not a nice, coherent entity, but rather a set of patches).

In response to various problems I have encountered, I have

corrected a few errors in the Fortran code;
created some additional Fortran utilities;
made changes to the Kincaid data set as distributed by John Irwin. In February 1999, a new "official" verison of John Irwin's data set has made my version obsolete; however, beware of the treatment of outliers in the "official" data set.

The details of this are explained in the present set of web pages.

An overview of the entire procedure I have used for my investigations is given in the note "Revised method compared to that of John Irwin".

All of the material I have described here is available as the package ASTM_SUP.ZIP (60 K, last update March 19, 1999).

For a broader overview, see the web page ASTM Standard Practice: Overview of materials.

Errors in the Fortran code

There was an error in John Irwin's program NEWFIT.FOR (version of October 24, 1997) , implying that only runs with the IPLT=6 make sense. This error has been corrected in a more recent version of the program, dated February, 1999.
It is also corrected in my revised version of NEWFIT.FOR, called NF.FOR.

Further, the NEWFIT program does not work as intended if IPLT=3 or 4. This error has also been corrected in my NF.FOR.

See below for details.

Additional Fortran utilities

I have created a slightly revised version of NEWFIT.FOR (called NF.FOR) which corrects the above mentioned errors, and which adds a few extra capabilities to NEWFIT. Overview of changes to NEWFIT.

Furthermore, I have written some extra programmes. You will find an explanation of what these programmes do and why they do it in the note "Revised method compared to that of John Irwin".

Changes to the Kincaid data set

There are many duplicate occurrences of arcs in the Kincaid data which was originally provided by John Irwin (dated April 10, 1997 and distributed prior to February 1999). Furthermore, that data set contained some outliers. In response to this problem I (HRO) created (in Spring 1998) a new version of John Irwin’s KININPUT.BAK data set where I removed many - but not all - duplicate arcs. Furthermore, I removed 5 outliers in order to be consistent with the Model Validation Kit. (See the file OUTLIERS.DAT in the Model Validation Kit for details). This revised data set was available off the web.

However, later (in February 1999) John Irwin has made a corrected data set available from his FTP site as part of the EVAL.ZIP package and it should now be regarded as the "official" version of J.I.'s data. The corrected data set is dated October 23, 1997. Note that outliers are marked as negative concentration values. This can be tricky if you use your own software to process the data (whereas John Irwin's NEWFIT behaves as it should). There is more information on the outliers here. The web page ASTM Standard Practice: Overview of materials provides information concerning availability of the EVAL.ZIP package.

The Model Validation Kit

Homepage of "Harmonisation..." initiative

This page is maintained by Helge Rørdam Olesen

Document date: May 25, 2001 - new pre-ample November 14, 2003.

Department of Atmospheric Environment, NERI