Generation of glacier IDs for GLIMS: perl code

[Bruce Raup]

Hi Frank,

> Just a short note on glacier IDs. For the new Swiss glacier inventory
> I have used a hydrological code (similar to the WGI code) for glacier
> identification, but converted to a pure number string without letters.
> I think the longitude / latitude code is ok for glacier identification,
> but I didn't know if a mixed letter/number code is easy to handle during

This is a good point, although I think it's a question of efficiency,
rather than possibility.  The idea was to have the IDs fairly
human-readable as well.

By the way, there have been several "votes" for glacier IDs using East
longitudes only.  Thus, the current suggested format is:

  GxxxxxxEyyyyy[N|S]

where xxxxxx is the East longitude, yyyyy is the north or south latitude,
and both have three digits right of the implied decimal point.  The E is
there only for human readability.  Anyone have further comments?

NSIDC's glacier would be "G254747E40013N".

> I actually have problems with the time stamp. How to
> identify glaciers that split up into parts during time by the code?
> This is perhaps not a problem for a 'first' inventory, but for the
> following ones or if a former inventory is available. Apart from glacio-
> logical questions, is there a strategy to track glacier parts by an
> identification scheme ?

The glacier ID identifies the glacier for all time.  For individual
analyses -- that is, a snapshot of a particular glacier at a specific time
-- there is a unique "analysis ID", and a timestamp.  If only part of the
glacier is visible for a particular year, then that's okay.  The polygon
in that case may have segments that bridge over clouds, say, and are
therefore "arbitrary".

Or maybe I've answered the wrong question.  Are you talking about glaciers
with a branching structure that, due to retreat, break up into more than
one glacier?  In that case, the new parts (or some subset of them) would
get new glacier IDs (primary key to the Glacier_Static table), and the old
ID would fill the "parent_icemass_id"  field of the new records.

See http://www.GLIMS.org/db_design.html.

Cheers,
Bruce

-- 
Bruce Raup
National Snow and Ice Data Center                     Phone:  303-492-8814
University of Colorado, 449 UCB                       Fax:    303-492-2468
Boulder, CO  80309-0449                                    braup at nsidc.org