<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
If the institution were not in the directory structure, and the same
model were run by several institutions, then they would have to be
careful to coordinate realization numbers and the rest of the
ensemble facet. These numbers aren't (and probably can't be)
standardized - they describe a wide range of choices made by whoever
runs the model.<br>
<br>
- Jeff<br>
<br>
On 3/13/13 5:52 PM, Karl Taylor wrote:
<blockquote cite="mid:51411F42.3090208@llnl.gov" type="cite">
<meta http-equiv="Content-Type" content="text/html;
charset=ISO-8859-1">
<font face="Times New Roman">Dear Galia and all,<br>
<br>
I'm glad thought is being given to unify directory structures
across projects. <br>
<br>
That being said it may not be known by everyone that:<br>
<br>
1) From the user's perspective this isn't really that necessary,
since the directory structure is hidden by ESGF (but see below
for further discussion).<br>
2) The ESGF publishing software doesn't care much at all about
the directory structure.<br>
3) In particular the ESGF search categories (facets) are not
tied to the directory structure.<br>
<br>
So there is quite a bit of flexibility allowed by ESGF, and
users can normally access data blissfully unaware of how its
been organized on various data nodes.<br>
<br>
Nevertheless, I agree that at least for some "projects"
organizing data in unified directory structures is useful. Node
managers who often want to access data outside of the ESGF api
can more easily find what they're looking for with a
standardized directory structure. If the directory structure is
rigorously enforced, users could take wget scripts created to
download one variable and easily modify it to download a
different variable (although its hard to be sure of success).<br>
<br>
Concerning your proposed directory structure, I have two
comments:<br>
<br>
1) I would recommend omitting the "institution" subdirectory.
For CMIP5 I think it was a mistake to include this (and also, I
wouldn't include it at a search category). The institution can
be recorded in the metadata of each file. If the same model was
run at more than 1 institution, then I'd like to see all the
simulations under a single "model" subdirectory, not split
across two directories under different institutions.<br>
<br>
2) When possible, I would stick with the terminology established
in the "CMIP5 Data Reference Syntax (DRS) and Controlled
Vocabularies" document available at: <a moz-do-not-send="true"
class="moz-txt-link-freetext"
href="http://cmip-pcmdi.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf">http://cmip-pcmdi.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf</a><br>
<br>
Best regards,<br>
Karl<br>
<br>
<br>
<br>
<br>
</font>
<div class="moz-cite-prefix">On 3/12/13 10:14 AM, Galia Guentchev
wrote:<br>
</div>
<blockquote cite="mid:513F6287.6050708@ucar.edu" type="cite"> Hi
everybody,<br>
<br>
Several groups have expressed interest to publish downscaled
climate datasets on ESGF. A standardized solution to publishing
(directory structure elements) would contribute to the prompt
identification of datasets. To discuss needs and options for
directory structure elements we had an initial teleconference
about a month ago. With this email we are expanding our reach to
other groups, such as the go-essp group, in order to have a
wider discussion of these elements.<br>
<br>
As agreed during our first teleconference, Aparna and Galia
worked on a proposal for a Directory Structure for publishing
downscaled datasets on ESGF. We would like to focus our next
teleconference on discussing this proposal. Below please find a
doodle poll for a potential next teleconference.<br>
<br>
<pre wrap=""><a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://doodle.com/hrwthqs2g5pgsyv6">http://doodle.com/hrwthqs2g5pgsyv6</a></pre>
<br>
**********************************************************************<br>
Details of each element of the proposed directory structure:<br>
<br>
Proposed elements -<br>
/projectID/<font color="#3333ff">sub-project</font>/product/institution/<b>predictorModel/experimentID/frequency/realm/MIPtable/Pred</b><b><br>
</b><b>ictor_experiment_rip/predictorversion</b>/<i>downscalingMethod/predictand
(variableName)/region</i>/<i>DownscaledDataversion</i>/file_name.nc<br>
<br>
Example:<br>
<br>
/ncpp2013/<font color="#3333ff">perfectModel</font>/downscaled/NOAA-GFDL/<b>GFDL-HIRAM-C360-coarsened/amip/day/atmos/day/r1i1p1/v20121024</b>/<i>GFDL-ARRMv1/tasmax/US48/v20120227</i>/tasmax_day_amip_r1i1p1_downscaled_US48_GFDLARRMv1_19790101-19831231.nc<br>
<br>
The new element <font color="#3333ff">sub-project </font>(in
blue above) gives the opportunity to indicate to users that in
the one case the method was trained on observations (standard
setting), and in the other on model that was considered to be
the truth (perfect model setting);<br>
The options there could be: PerfectModel or Standard - where
possibly there could be a different name instead of 'standard'
for the standard downscaling setting.<br>
<br>
<p>For NASA datasets some of the directories could be: <br>
</p>
project = NEX<br>
product = downscaled<br>
institution = NASA-Ames<br>
predictorModel - original model value<br>
experimentID = historical<br>
frequency = mon<br>
realm = atmos<br>
Predictor_experiment_rip - original model value<br>
variable = precipitation or temperature<br>
region = CONUS<br>
<br>
DownscalingMethod will also be included as a directory to allow
for search on method.<br>
<br>
**********************<br>
There are a set of sub-directories that refer to the <u>PredictorModel</u>
- presented in bold - <b>/predictorModel/experimentID/frequency/realm/MIPtable/Pred</b><b><br>
</b><b>ictor_experiment_rip/predictorversion</b><br>
<br>
Where: <br>
<ul>
<li>predictor model - is the specific GCM which is the source
of the predictor data set - GFDL-HIRAM-C360-coarsened - in
the above example</li>
<li>experimentID - the specific experiment - amip in this case</li>
<li>frequency - refers to the temporal scale of the predictor
fields - daily</li>
<li>realm - the realm of the predictors - in this case
atmos(phere)</li>
<li>MIPtable - name of the model intercomparison table - daily
in this example, could be amon - for atm monthly data;</li>
<li>Predictor-Experiment-rip - follows the standard notation
from CMIP5</li>
<li>version - the version date of the global model that
provided the predictor dataset</li>
</ul>
<p>The elements above follow quite closely the structure for
CMIP5 model output directory elements.<br>
</p>
There is a set of sub-directories that refer to the Downscaling
method - presented in italics - <br>
<i><i>downscalingMethod/predictand (variableName)/region</i>/<i>DownscaledDataversion<br>
</i> <br>
</i>Where:<br>
<ul>
<li>downscalingMethod - is the downscaling method abbreviation
- in this case GFDL-ARRMv1 - the GFDL in the name indicates
that this is a setting applied by GFDL where there were two
sets of predictors, based on the ARRM method of K.Hayhoe;
also v.1 indicates which version of the ARRM method was used
(the original version) - more details about the method are
given in the global attributes of the file;</li>
<li>Predictand (variableName) - the specific predictand
variable that was downscaled; tasmax in this case;<br>
</li>
<li>region - indicates that the method was applied to the US48</li>
<li>DownscaledDataversion - the version of the downscaled
dataset<br>
</li>
</ul>
<p><b>For the purposes of standardization there are two
directions to consider:</b> <br>
</p>
<p>1) One is to have<b> one standard directory</b> structure
that will be used by all - for example, following the example
of GFDL to have the details of the predictor model first and
then the downscaling method details:</p>
<ul>
<li>ProjectID - sub-project - product - Institution -
Predictor dataset details - Downscaling method details -
Filename</li>
</ul>
<p> Having a standardized approach would help any automated
service/web service to detect the directory path for a
particular dataset. </p>
<p>2) During our last teleconference there was a proposal to
follow the downscaling practice and describe the downscaling
method first and then the predictor model. This leads to <b>two
paths</b>:</p>
<p> • ProjectID - <u><font color="#3333ff">Standard or
Perfect Model sub-project facet</font></u><font
color="#3333ff"> </font>- product - Institution - then see
below:<br>
- (if Perfect model setting) Predictor dataset
details - Downscaling method details, <br>
- (if Standard setting) - Downscaling method
details - Predictor dataset details<br>
</p>
<p><br>
The NCPP Core team accepts that it may be reasonable to have a
directory structure - where the method description is first;
and another directory structure - where the predictor
description is first and then the methods that are applied are
described; <b>NCPP will support either approach</b> (one
overall directory structure, or two separate pathways) and if
the second approach is chosen (with two different
sub-directory sequences) - we would like to promote and to
support the standardization of these different directory
pathways - meaning - we will support two standardized
directory structures to accommodate two common practices.<br>
</p>
<p><br>
</p>
******************<br>
<font color="#009900">Additional details<font color="#000000">:
</font></font><br>
<br>
<b>Variable level attributes-</b><br>
The published dataset should also conform to CF-standards. <br>
eg-<br>
<br>
tasmax:long_name = "Downscaled Daily Maximum
Near-Surface Air Temperature"
;
<br>
tasmax:units = "K"
;
<br>
tasmax:missing_value = 1.e+20f
;
<br>
tasmax:_FillValue = 1.e+20f
;
<br>
tasmax:standard_name = "air_temperature"
;
<br>
tasmax:original_units = "K"
;
<br>
<b> tasmax:downscaling_method: GFDL-ARRMv1</b><br>
<br>
<b>Global attributes- </b>listing a few here, several
CMIP-style attributes will be inherited. <br>
<br>
"predictorModel" will replace "model_id"<br>
For the 'downscaling model', as agreed with Luca on the call
it would be 'downscalingMethod' <br>
<br>
:Conventions = "CF-1.4" ;<br>
:references = "info about model, training
datasets etc will be provided here"<br>
:info = "additional info about the downscaling
method" <br>
:creation_date = "2011-08-19T21:57:06Z" ;<br>
:institution = "NOAA GFDL(201 Forrestal Rd,
Princeton, NJ, 08540)" ;<br>
:history = "info on file processing. Eg"
processed by toolX." ;<br>
:projectID = ncpp2013<br>
:subprojectID = perfectModel<br>
:product = downscaled<br>
:institution = NOAA-GFDL<br>
:predictorModel = GFDL-HIRAM-C360-coarsened<br>
:experimentID = amip<br>
:frequency = day<br>
:modeling_realm = atmos<br>
:Predictor_experiment_rip = r1i1p1<br>
:region = US48<br>
:table_id = day<br>
:version = v20120227<br>
:downscalingMethod = GFDL-ARRMv1<br>
**************************************************<br>
<br>
Best regards,<br>
Galia and Aparna<br>
<br>
<pre class="moz-signature" cols="72">--
Galia Guentchev, PhD
Project Scientist
National CLimate
Predictions and
Projections
Platform (NCPP)
NCAR RAL CSAP
FL2 3103
3450 Mitchell Lane
Boulder, CO, 80301
phone: 303 497 2743 </pre>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
GO-ESSP-TECH mailing list
<a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:GO-ESSP-TECH@ucar.edu">GO-ESSP-TECH@ucar.edu</a>
<a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://mailman.ucar.edu/mailman/listinfo/go-essp-tech">http://mailman.ucar.edu/mailman/listinfo/go-essp-tech</a>
</pre>
</blockquote>
<br>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
GO-ESSP-TECH mailing list
<a class="moz-txt-link-abbreviated" href="mailto:GO-ESSP-TECH@ucar.edu">GO-ESSP-TECH@ucar.edu</a>
<a class="moz-txt-link-freetext" href="http://mailman.ucar.edu/mailman/listinfo/go-essp-tech">http://mailman.ucar.edu/mailman/listinfo/go-essp-tech</a>
</pre>
</blockquote>
</body>
</html>