Changed in GISTEMP

[Nick Barnes]

I'm looking at updating ccc-gistemp to reflect changes in GISTEMP.
These are my ill-formed notes.

1. GHCNv3 replaces GHCN and USHCN.

Happily we already have GHCNv3 code (thanks to David).
STEP0/step0_README.txt gives these instructions:

ghcnm.tavg.latest.qca.tar.gz (from GHCN v3) - type: "gunzip *gz" and "tar -xpf *.tar"
then: ln ghcnm.v3.1.0.*/*.dat adj_data

2. SCAR records.

We should scrape these from the BAS website, rather than from the
GISTEMP sources. It looks pretty straight-forward. We could host the
station ID files (that we currently get from the GISTEMP sources)
somewhere ourselves.

3. Ts.strange. adjustment files

STEP0/input_files/Ts.strange.RSU.list.IN_full is now called
STEP0/input_files/Ts.strange.v3.list.IN_full

and has far fewer entries, although it has the same format.

4. STEP1

STEP1 loses Ts.discont.RS.alter.IN and combine_pieces_helena.in

The comb_records, comb_pieces, and alter_discont steps are completely
omitted: the step 1 pipeline used to be:

v2_to_bdb | comb_records | comb_pieces | drop_strange | alter_discont | bdb_to_text

now it is:

v3_to_bdb | drop_strange | bdb_to_text

(note that v3_to_bdb is identical to v2_to_bdb apart from filenames).

5. STEP2

This is the only semantically interesting chnge in step 2:

+ if(LightGl>2) read(li(108:109),*) indexGL ! Berkley Earth data
name(31:31)='1' ! rural
+ if(LightGl>1.and.indexGL>0) name(31:31)='2' ! rural=pitch dark

6. STEP 3

STEP 3 is extended by adding mid-latitude zones (23.6 to 64.2
degrees), combining the second and third latitude bands.

7. STEP4_5

Again, the mid-latitude zones. Also:

diff -ur ../2010-09-10/GISTEMP_sources/GISTEMP_sources/STEP4_5/convert1.HadR2_mod4.f GISTEMPv3_sources/STEP4_5/conve\
rt1.HadR2_mod4.f
--- ../2010-09-10/GISTEMP_sources/GISTEMP_sources/STEP4_5/convert1.HadR2_mod4.f 2007-09-07 14:47:17.000000000 +0100
+++ GISTEMPv3_sources/STEP4_5/convert1.HadR2_mod4.f 2011-02-02 19:57:00.000000000 +0000
@@ -45,6 +45,17 @@
read(14) line,clim
close (14)
c
+c Skip some regions where SST is impacted by nearby ice floats
+c
+ do j=166,180
+ do i=1,360
+ do m=1,12
+ clim(i,j,m)=bad
+ end do
+ end do
+ end do
+
+c
c Read in the SST data for recent years (unit 11)
c
call getarg(1,line)

But that seems to be it. There are no code changes corresponding to
the shift from Hadley to ERSST datasets; presumably this is reflected
in a changed SBBX.HadR2.

8. STEP6

There are some changes in STEP6, but we don't really deal with that
anyway.

[Nick Barnes]

On Wed, May 23, 2012 at 3:52 PM, Nick Barnes <n...@ravenbrook.com> wrote:

4. STEP1

STEP1 loses Ts.discont.RS.alter.IN and combine_pieces_helena.in

The comb_records, comb_pieces, and alter_discont steps are completely
omitted: the step 1 pipeline used to be:

v2_to_bdb | comb_records | comb_pieces | drop_strange | alter_discont | bdb_to_text

now it is:

v3_to_bdb | drop_strange | bdb_to_text

(note that v3_to_bdb is identical to v2_to_bdb apart from filenames).

Also, and this is surely related to the change from v2.inv to v3.inv that I noted earlier, the ushcn2.tbl file is replaced by ushcn3.tbl.  The latter is mechanically produced from the former by ushcn3.tbl.f, which produces 11-digit station IDs by simply prepending 425000 to the 5-digit USHCN id.

--
Nick Barnes, Climate Code Foundation, http://climatecode.org/