[Go-essp-tech] discrepancies among gateways

[Karl Taylor]

Dear all,

I noticed that as of 15:30 PDT on 7/22/11 that different ESG gateways 
are serving different amounts of CMIP5 data:

gateway           # of  CMIP5 datasets available
----------            -----------------------------------
PCMDI                 6973
BADC                   6973
DKRZ                   6973
NCI                       6974  (why 1 more than the above????)
JPL                        6457  (missing CSIRO-MK3-3 model)
NCAR                   6457 (missing CSIRO-MK3-3 model)
NERSC                 6458 (missing CSIRO-MK3-3 model)
ORNL                         0

I noticed several other problems:

1.  NERSC gateway:  the search categories are not correctly populated.

2.  NCI gateway:  If you click on CMIP5 in search category "project" you 
get an error message instead of a list of datasets.
Error: Failed to convert property value of type [java.lang.String] to 
required type [int] for property limit; nested exception is 
java.lang.NumberFormatException: For input string: ""
If you then click on CSIRO-Mk3-6-0 in the "model" category, you get 516 
datasets returned, which is proper behavior.  Curiously, if you then 
click on "any model" you get the complete CMIP5 list of datasets, as you 
should have when you previously got the error message.

3.  NCAR gateway:  CMIP5/IPCC AR5 appears as one of the "projects" 
(along with "cmip5").  This is confusing.  We want users seeking CMIP5 
data to go to cmip5 and not get distracted by CMIP5/IPCC AR5.  Please 
remove project CMIP5/IPCC AR5.

4.  There are inconsistencies among the gateways in rendering the list 
of models that appear under the "model" search category.  A common 
problem is that the word "model" appears before some of the model names 
on some of the gateways (but not others).  Also there are other minor 
differences (e.g., lower case in the place of uppercase acronyms).  We 
want users to get used to the "official" CMIP5 names, which they should 
use in publications.  These official names can differ slightly from the 
names that appear in the filenames.  The official model name is supposed 
to have been recorded in each file as a global attribute called 
"model_id".  From information sent to me my the groups, following 
official model names should appear in the search list (or if this is not 
possible, the name of the model as recorded in each filename should be 
used (sorry that sometimes more than 1 model appears in a table cell):

Official model_id             model name appearing in filename
---------------------              
------------------------------------------
ACCESS 	ACCESS
BCC-CSM1.1 	bcc-csm1-1
BNU-ESM 	BNU-ESM
CanESM2, CanCM4, CanAM4 	CanESM2, CanCM4, CanAM4
CCSM4 	CCSM4
CCSM4(RSMAS) 	CCSM4-RSMAS
CESM1(BGC) 	CESM1-BGC
CESM1(CAM5) 	CESM1-CAM5
CESM1(CHEM,CAM5) 	CESM1-CHEM-CAM5
CESM1(CHEM) 	CESM1-CHEM
CESM1(WACCM) 	CESM1-WACCM
CMCC-CESM 	CMCC-CESM
CMCC-CM 	CMCC-CM
CMCC-CMS 	CMCC-CMS
CNRM-CM5 	CNRM-CM5
CSIRO-Mk3.6 	CSIRO-Mk3-6
EC-EARTH 	EC-EARTH
FGOALS-G2.0 	FGOALS-G2
FGOALS-gl 	FGOALS-gl
FGOALS-S2.0 	FGOALS-S2
FIO-ESM 	FIO-ESM
GEOS-5 	GEOS-5
GFDL-CM2.1 	GFDL-CM2P1
GFDL-CM3 	GFDL-CM3
GFDL-ESM2G 	GFDL-ESM2G
GFDL-ESM2M 	GFDL-ESM2M
GFDL-HIRAM-C180 	GFDL-HIRAM-C180
GFDL-HIRAM-C360 	GFDL-HIRAM-C360
GISS-E2-H 	GISS-E2-H
GISS-E2-H-CC 	GISS-E2-H-CC
GISS-E2-R 	GISS-E2-R
GISS-E2-R-CC 	GISS-E2-R-CC
GISS-E2CS-H 	GISS-E2CS-H
GISS-E2CS-R 	GISS-E2CS-R
HadCM3 	HadCM3
HadCM3Q 	HadCM3Q
HadGEM2-AO, HadGEM2-A
	HadGEM2-AO, HadGEM2-A
HadGEM2-CC 	HadGEM2-CC
HadGEM2-ES 	HadGEM2-ES
HiGEM1.2 	HiGEM1-2
INM-CM4 	inmcm4
IPSL-CM5A-LR 	IPSL-CM5A-LR
IPSL-CM5A-MR 	IPSL-CM5A-MR
IPSL-CM5B 	IPSL-CM5B
MIROC-ESM 	MIROC-ESM
MIROC-ESM-CHEM 	MIROC-ESM-CHEM
MIROC4h 	MIROC4h
MIROC4m 	MIROC4m
MIROC5 	MIROC5
MPI-ESM-HR 	MPI-ESM-HR
MPI-ESM-LR 	MPI-ESM-LR
MPI-ESM-MR 	MPI-ESM-MR
MRI-AGCM3.2H 	MRI-AGCM3.2H
MRI-AGCM3.2S 	MRI-AGCM3-2S
MRI-CGCM3 	MRI-CGCM3
MRI-ESM1 	MRI-ESM1
NorESM-M, NorESM-ME, NorESM1-L 	NorESM-M, NorESM-ME, NorESM1-L


I have not confirmed that model_id is necessarily consistent with the 
names in the first column, nor that the filenames are consistent with 
the 2nd column.  If there are inconsistencies, I would like to hear 
about them.

Because of the problems above (other than 4), the CMIP5 website now 
(temporarily) directs researchers to only the PCMDI, BADC, and DKRZ 
gateways.  Only at these sites is both the search engine working o.k. 
and all the output from CMIP5 can be seen.  We don't want to direct 
users to sites where they will encounter problems or be unable to see 
some of the CMIP5 output.

Best regards,
Karl






-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/go-essp-tech/attachments/20110722/891873b1/attachment.html