[kepler-dev] Fwd: ENM workflow + a quick Kepler tutorial
Deana D. Pennington
dpennington at lternet.edu
Fri Mar 18 06:18:57 PST 2005
Iagree; I think these problems are SMS problems, although seems like the initial
query should be able to exclude points without lat/long information. We do need
to integrate Taxon in that we need to be able to pass the query results to them
for name resolution, then get their results back into Kepler.
Dan: I think Jianting and I are going to put some effort over the next couple of
months into the post-GARP part of the workflow (best subsets, etc). If you
could send me whatever notes you took in Dec, I'd like to compare what he told
you to what he told us in Feb, so I can make sure we get it the way he wants it.
Quoting Shawn Bowers <sbowers at ucdavis.edu>:
> Aren't most of the highlighted issues EML or even semantic typing
> issues, i.e., that a data set has lat/lon values or that a data set
> contains species abundance/presence/biomass measurements tied to a
> location (that can be translated somehow to a lat/lon)?
> I think A, B, C, and D could all be handled via semantic mediation
> (e.g., see our SWDB04 paper). Of course, this work is not implemented
> within Kepler, but will be soon, and could be done even with very simple
> ontologies such as the one described in that paper (note that the
> prototype in the paper was implemented).
> >> Current Status of ENM (GARP) workflow - 17 Mar 2005
> >> *1) Species occurrence information
> >> In current examples, a search has been made on the 'Data' tab for
> >> 'Mephitis' [a skunk]. There are multiple hits (~32), but only 2 of
> >> these have lat, long location data. This is determined by a visual
> >> inspection of the Digir data. The 2 data sets are concatenated using
> >> Kepler expression actor and the '+' operator. *
> >> *There are several problems with the current approach:
> >> A) Searches return multiple hits and there is no automatic way to
> >> determine which species searches have location data.
> >> B) Given results with location data, we do not have a way to
> >> automatically combine the results into a (long, lat) list.
> >> C) Somehow, the Taxon group work needs to be integrated into the
> >> locations tabulation. HOW DO WE DO THAT?
> >> D) Does sufficient location information exist yet for species of
> >> interest? What are the species of most interest? [We need a number of
> >> locations to subset for training to build the best subsets workflow.]
> >> 2) Climate/Environment layer preparation
> >> 3) Best subset calculation
Dr. Deana D. Pennington
Long-term Ecological Research Network Office
UNM Biology Department
1 University of New Mexico
Albuquerque, NM 87131-0001
505 272-7080 (fax)
More information about the Kepler-dev