26.09.2013 BiolDiv

Meeting 26 September 2013, 12 noon

Google Hangout

Present: Anton, Lino, GP, Fabio, Edward

Notes

Species Distribution Modelling

Time is getting very short to do anything with the improved resolution data, and with openModeller, as it was originally planned. OpenModeller is now available in the OpenBio infrastructure, but limited to the coarse-resolution environmental data that are distributed together with the desktop version of openModeller, so not suitable for our purposes. As an alternative, the R package Dismo or the Java library Weka could be considered. Environmental enrichment plans also have not come to fruition. As an alternative, we could work for the time being with a combination of BioOracle and GEBCO (or ETOPO) data.

As an interesting exercise which also has scientific value, we might look into modeling of cod data. Cod is monitored in several fisheries bodies; survey data are looking typically at a limited number of species. Absence of a species in a data/time in the dataset of that survey can be inferred as an absence. The exercise would be to see how these inferred absences improve our species distribution models. Edward will write up a small document; further work on this exercise will e discussed, including which tools to use (e.g. Weka vs Dismo).

Bio-oracle data are only distributed as ascii layers, and some of the information (min-max lat-lon of the raster, projection…) are missing. Need to further explore how to transform into rasters that are easy to use in the infrastructure (e.g. NetCDF).

Environmental data enrichment

For the 4d interpolation needed for environmental enrichment, many use Diva GIS. But still needs expert knowledge to interpret autocorrelation in different dimensions… So a discussion with NODC experts is still very valuable. Early next week we’ll have a brief conference call between Fabrice, Julien, GP and Edward; this call will be specifically to prepare for a conference call with NODC people (note that the latter call still has to be suggested to NODC; up to now the only contacts have been through email). Nicolas’ involvement will be very useful – both for his expertise as ichthyologist, and his experience with AquaMaps.

BiOnym

Abstract on BiOnym was submitted for the TDWG meeting. Edward would love to attend, but doesn’t have funds (and the registration fee is prohibitively high to do this on a personal basis). A request for funding was sent to Donatella and Marc. Lino and Anton will assist with follow up.

Fabio is working on a separation of the specimen.jar and the reference lists it needs. Darwin Core Archive is a first step, but this has to be expanded with some calculated fields. Scientific names are priority at the moment, but vernaculars will be taken on board right now as well, as this will not really impact the work load for Fabio. ASFIS will soon be available as DwCA as well, provided by FAO, and will be taken on board.

Two approaches of Casey have been made available on the iMarine infrastructure through the statistical manager. Same could be done with specimen.jar.

Edward is working on a prototype for BiOnym in R; the prototype relies heavily on calls to specimen.jar, and on PostgreSQL. The aim is to show how different parts of the taxon name matching activities could be pulled together in on flexible workflow approach.

Next meeting

Will be held 3 October 2013, 11 am (!! one hour earlier!!). Lino will call the meeting to order; Edward will send a reminder earlier that day.