These are dated November 14th, and seem to still be work in progress –

http://data.giss.nasa.gov/gistemp/sources/GISTEMP_sources.tar.gz , linked from the Sources page, seems to have become a broken link for now.

While http://data.giss.nasa.gov/gistemp/updates/ has been updated to indicate that USHCN version 2 will be used from November 13th, there is as yet no mention of STEP6 there, and it is not mentioned in the revised gistemp.txt file either.

A quick look at the STEP6 files indicates that it produces “line plots” from ANNZON.Ts.GHCN.CL.PA.1200, ANNZON.Tsho2.GHCN.CL.PA.1200, ZON.Ts.GHCN.CL.PA.1200 and ZON.Tsho2.GHCN.CL.PA.1200.

All you need is a Linux box (I’m on a Red Hat 7.2 release on an AMD 400 mhz chip with about 128 MB of memory IIRC – you don’t need a lot of hardware. It started life as a 486 machine a couple of decades ago; I wanted to be “period correct” (That’s a re-enactor inside joke The changes needed to the code are not particularly large (though a bit hard to discover) mostly related to not using a non-standard extension that lets you put data initialization into the variable declarations. (i.e. use a DATA statement to load initial values…)

That, and some programs need an f77 compiler while others need f90 or better (i.e. g95). I have a makefile that keeps that straight in my build. Oh, and I yanked out the ksh directives and made it sh / bash compatible (mostly adding a couple of strategic “;” characters that are optional with ksh… )

I used the free g95 compiler for my “f90 or better” and it seems to work well.

The only “issue” still open for me is that the Hadley SST file is “bidgendian” and the PC I’m on is “littleendian” so I can compile, but not run, STEP3 and STEP4_5. I used the stable.91 g95 and the newer release claims to support the “convert=bigendian” flag that would solve this problem.

If anyone wants a “tarball”, just leave a note on one of my blog pages or send email to the address in my “about” tab; and I’ll put it on your FTP destination. I can also put up a set of “diffs” if folks would rather have that. FWIW, I’ve also “cleaned things up a bit” by moving to a better directory structure and naming things a bit more cleanly (shell scripts now end with .sh on a reliable basis; binaries are in a “bin” directory and sources in a “src” directory). I also pulled the “inline compile and run” out and made a “makefile” See:

http://chiefio.wordpress.com/2009/07/29/gistemp-a-cleaner-approach/

The direcotory change is not needed to run, it just makes it a lot easier to keep straight what’s what.

Oh, and if you put this on a bigendian box you ought to be “good to go” through STEP4_5 as is. (Mac PowerPC, SUN Sparc, IBM RS6000, MIPS and more are “bigendian”. Intel x86 are littleendian. HP made a box with selectable endian character at one point).

Basically, if anyone wants this on their box, I’m available and willing to help. If you are “L.A. close” to San Franciso, I could be talked into assisting (gas money would be nice; “will program for beer”
E. M. Smith

http://chiefio.wordpress.com/

he relates his observations on the data and how much of the warming seems related to stations entering and leaving the GISS records.

On or about 7 September 2007, Dr. Hansen said…

People interested in science may want to wait a week or two for a simplified version.

I am interested in science and I have now waited a week or two (or somewhat longer) for a simplified version. Can anyone tell me where I might find it?

3) System requirements

The set-up code and automatic compilation rely heavily on gmake (version 3.79 or higher) and perl (version 5.005 or higher). Both of these programs are public domain and available for almost all platforms. Up-to-date versions are required to be able to set up and compile the model.

The code requires a FORTRAN 90/95 compiler. It has been tested with the SGI, IBM, and COMPAQ/DEC compilers on their respective workstations. For Linux, Macs or PCs, the choice of compiler is wider, and we have not been able to test all possibilities. The Absoft ProFortran compiler works well, as do the Lahey/Fujitsu and Portland Group compilers. We have been unable (as yet) to get the model to compile with the Intel or VAST compilers due to internal errors in those products. Please let us know if you succeed (or more importantly, fail) in compiling the code using any other compiler.

Note that the parallelization used in the code is based on the OpenMP shared memory architecture…….