[Seavox] "Sea Areas"

Sat Nov 1 09:28:32 GMT 2008

Hi Olivier,

To me the killer arguments to adding the term 'regional' to the new Sea Area vocabulary are:

(1) It cannot be related to a spatial object.
(2) I cannot think of a way to link it into the concept map hierarchy

There is no doubt that you have a valid issue - identical discovery metadata for different datasets clearly shows this, but it's a shortcoming in our metadata model that should not be solved by shoe-horning ill-fitting terms into vocabularies.

My view is that we need an additional optional keyword, but your problem and its solution is something that should be discussed and resolved by the SeaDataNet TTT.

Cheers, Roy.

>>> "olivier lauret" <olauret at cls.fr> 10/31/08 3:45 PM >>>
-----Message d'origine-----
De : Roy Lowry [mailto:rkl at bodc.ac.uk] 
Envoyé : lundi 27 octobre 2008 22:18
À : seavox at biwebs1-2.nerc-liv.ac.uk; olivier lauret
Objet : RE: [Seavox] "Sea Areas"

Hi Olivier,

First, I have no objection to adding 'Pacific Ocean' and 'Atlantic Ocean' into the hierarchy unless anybody can raise arguments against.

àOK, thank you Roy.

Secondly, I strongly feel that dataset resolution is a totally separate issue.  With the proposed vocabulary I am trying to address 'spatial coverage' in the ISO19115 domain, not dataset resolution. 

à MMmmmmh..Not exactly. Of course I totally agree that your vocabulary server is about ISO 19115 location, but what I was talking about is not properly a 'resolution' information, I think. Besides, this is exactly the problem: the words "global" and "regional" can mean [for remote sensing & models users] "coverage"; for you it means "resolution". Actually to me in that case it is neither a coverage, nor a resolution, but probably something at the intersection.

I have no problems at all with your vocabulary terms for resolution that could be used to populate a relevant field in an ISO19115 profile.  The problem is that our current SeaDataNet ISO19115 profiles don't have suitable extensions for this information.  This is something we in the TTT need to look at for the next version of the SDN profiles.  For now, the most important issue for me is that we don't pollute the spatial coverage list I plan to build with terms that use the same words for a totally different entity.

àYes extending SDN ISO19115 profile in a next version to take into account such information, or such a way to understand information, could be a good thing. It is typically the kind of domain specific information that cannot be supported by default by ISO 19115, which is a generic shell.

It is far from my mind to pollute the vocabulary server, I just want to get your opinion about this question: when proposing a dataset on a given area through a database, a search engine, or through a mapping software, what can distinguish two 'twins' datasets with the same parameters?

An example. A few months ago we discussed with people involved in JRA5 (Mediterranean Sea) about climatologies. They were asking if the data used to compute the climatologies were geographically extracted from the usual "global" dataset, or coming from the specific "Mediterranean Sea" dataset. 

In that case two different datasets can be described by the same metadata, they are both covering the same area, which is Mediterranean Sea. We use to face this problem, actually, and the only solution we found was to add the words "global" and "regional" in the dataset description. It is the case for instance in the MERSEA search engine (http://www.mersea.eu.org/Information/dynamic/MerseaSearch?profile=partner), that implements this concept.

Such "global" and "regional" terms wouldn't belong to the "Sea areas" vocabulary server. The only consequence I see on the vocabulary server  is that using "Sea areas" list would be mandatory only if it is a "regional" dataset.

Cheers,

Olivier.

Olivier Lauret

CLS - Space Oceanography Division

Products Dissemination Dpt

8-10 rue Hermes, 31520 Ramonville St.Agne

France

Tel. (+33) (0) 561 39 48 51

Fax:(+33) (0) 561 39 47 80

 	 http://www.cls.fr <http://www.cls.fr/> 
http://www.aviso.oceanobs.com/ <http://www.aviso.oceanobs.com/> 

http://www.esdi-humboldt.eu/home <http://www.esdi-humboldt.eu/home> 

http://www.seadatanet.org/ <http://www.seadatanet.org/> 

netCDF Climate and Forecast <http://cf-pcmdi.llnl.gov/> 

Cheers, Roy.

>>> "olivier lauret" <olauret at cls.fr> 10/22/08 5:32 PM >>>

Dear Roy,

Sorry for the delay, I'm back to my office after vacation. Completely agree with your plan, especially about your project of spatial database working with OGC webservices! I have a few comments that I'd like to live to your assessment:

- Areas such as Pacific Ocean and Atlantic Ocean are not provided; the fact is that it is important for ocean models and remote sensing data sets to have them as keywords. In other words I wish the ontology allowed a relationship like : 

World >> Pacific Ocean >> North Pacific Ocean >> ..

      >> Pacific Ocean >> South Pacific Ocean >> ..

- Still in the same field of view, ie about "remote sensing and model world". In these communities people use to speak about 'global' coverage versus a given area, for ex Mediterranean Sea. As 'global' is not properly an area but more a space scale, that means "worldwide" but not only, how do Seadatanet can deal with such important questions? 

Here at CLS we choosed to introduce a vocabulary keyword about those scales. If the keyword value is 'global' then nothing about geographic areas names. If the keyword value is 'regional' then it is mandatory to mention which one is the geographic area concerned, following Seavox. That's a possibility, but there must be something better to do.

I remember we already talk about that, but I am not sure of the status of this discussion?

What is your opinion on that?

Thanks!

Olivier.

-----Message d'origine-----

De : seavox-bounces at biwebs1.nerc-liv.ac.uk [mailto:seavox-bounces at biwebs1.nerc-liv.ac.uk] De la part de Roy Lowry

Envoyé : vendredi 10 octobre 2008 17:40

À : seavox at biwebs1-2.nerc-liv.ac.uk

Objet : [Seavox] "Sea Areas"

Dear All,

Attached is a concept map of a proposed nucleus list of "Sea Areas" (defined something like named subsets of the Earth's surface of interest to the oceanographic data management domain for use as data discovery keywords).  Consequently, it includes chunks of land to provide context for inland waters that are increasingly falling into the oceanographic data centre sphere of interest).

The terms included come from four sources as indicated by the bubble colour.  The pale blue terms bear a striking resemblance to the unpublished IHO 2002 S23 revision.  The red/orange terms are the so-called ROSCOP extensions to C23.  The green terms are terms I've added to cover, with geographical context, freshwater bodies  that have been brought to my attention as being of interest to 'oceanographic data centres'.  The purple term is the cause of all the controversy in IHO and the reason why S23 hasn't been revised since 1953.

The terms are organised into an ontology that may be navigated by starting at 'World' then following the arrows.

My plan would be to implement these terms as a single list in the vocabulary server with the relationships between terms as shown by the arrows.  Further terms may be added as required, providing they are proposed together with the terms to which they should be mapped (ie broader areas that enclose the new term or overlap with it).  This new list would replace the existing C16 vocabulary used in SeaDataNet. Naturally mappings to the new list would be provided wherever possible.

However, I plan to go further than this and set up a spatial database holding a polygon corresponding to each term that would be accessible through WFS and/or WMS.  Service URLs would be included in the term definition field.

Initially, I would like the governance group to consider the following questions:

(1) Are there any reasons why this plan shouldn't go ahead?

(2) Should 'Sea of Japan' be included as a term under SeaVoX governance?

(3) Should we stick with the local spellings such Jawa Sea and Tai-Wan Strait?  If not, which terms should we change?

Any other comments on the terms included would be welcome.  

Note that I would rather avoid requests for additional terms at this stage until I get something together.  

The concept map is also available in CMAP XML export format if anyone needs it.

Cheers, Roy.

                           Cliquez sur l'url suivante https://www.mailcontrol.com/sr/Kzk9kr2MHj7cohtw78djyWSR9TR7nOeiKUlp4!vWWpitISgtmjGvdZiQg7XMXKVHkcVX+qjV7oOzA9hj6iEC+Q==  

                    si ce message est indésirable (pourriel).

-- 

This message (and any attachments) is for the recipient only. NERC

is subject to the Freedom of Information Act 2000 and the contents

of this email and any reply you make may be disclosed by NERC unless

it is exempt from release under the Act. Any material supplied to

NERC may be stored in an electronic records management system.