Project

General

Profile

Home » Database development » Meetings » 2011-10-24 to 28 working group »

2011 working group Fr Summary

BIEN science

Range size spectra

  • a lot of similarities in Amazon and deciduous forest means and medians
  • skewness similar
  • sampling bias towards small range species in cell
  • sample size dist in cell
  • using range calculations to put patterns on map
  • species in 1 deg grid cell
  • species pool based on range overlap
  • is observed trait range different from random?
  • look at dataset over time
  • in dry areas, triggered by availability of water

Functional traits vs range

  • conservatism in traits for closely related species
  • range not conserved at all
  • absolute range size
  • model of branch lengths
  • auto-prune to BIEN species list
  • factor in missing taxa at internal nodes
  • commonness of rare species in BIEN
    • not properly cutting data to new world
  • species occurring along mountain areas and areas w/ high levels of animism
  • macro-eco relationships
  • patterns opposite those for birds and mammals
  • tight relationships btw seed size and plant size

Science interface

  • simclim: climates
  • phylo: phylogeny
  • env: temperature, pcpt
  • valid: family, taxon name
  • TNRS -> corrected species, family name, authority
  • annotate data with species list, lat/long
  • climate from lat/long server
    • needs password protection
  • list of points vs box
  • ensure no data leaks
  • geospatial meeting at 1:30pm

Database

  • goal is to have something to walk away from in 1 year
  • same validations but more data
  • standard schema VegX/DwC to load data
  • test VegX, VegBank to make sure they can accommodate use cases
  • test load of sample data into VegBank
  • reload data into core db
    • FIA data in FTP site
  • in place in Apr
  • load full db in Apr/May
  • use Nick Spencer's mapping tool to streamline the process of getting plot data into VegBIEN
  • validation scripts will run on core db
  • end access point to retrieve data
  • CSV reps of db and range maps need to be accessible
  • derived products
  • at end of 1 year, have sustainable model
  • how suitable BIEN 2 is w/ minor tweaks? vs. VegBank-based VegX-compliant model
  • whenever can't get something out of BIEN 2, let us know
  • need use cases that would break the system
    • asking more from the initial analysis
  • many manual processes before data is ready
  • part of wiki w/ data problem statements: bug list

Data problems

  • ability to separate different plot types: 10,000 Ha should be m^2: different units
  • data cleaning, not just scrubbing
  • engage TDWG (/Tadwig/)
  • missing element, specimen, or occurrence record
    • e.g. cultivated flag

Future plans

  • another meeting next year?
  • done after BIEN 3?
  • funding for more meetings?
    • NSF/RCN grant for 5 years support for meetings
  • informatics proposal from NCEAS
  • additional funding for tech devel
  • compelling scientific question that can't be addressed because of tech challenges
    • information plan complemented by what scientists working on
  • educate NSF about what we're doing
  • press release about what's being funded
  • taxonomies, phylogenies, functional traits
  • keep the BIEN server running
  • will data be out of date in 3-4 years?
  • BIEN vet
  • addition of iPlant as collaborator
  • bring GBIF in as partner
    • move beyond botany to other organisms?
  • map of life people focusing on vertebrates, but interested in what were doing
  • going global would broaden the scope
  • when does BIEN become accessible to the rest of the world? how to do it?
    • objections from data providers who want data private
  • new db will have tiered access control
  • who will scrub data?
  • a lot of work involved in going global?
  • instead of scrubbing, we already have globally distributed data
  • politics in going global
  • start geoscrubbing
  • internationalization in species names, place names
  • touch base in conference call before Thanksgiving
  • can put things together by Dec/Jan NSF deadline?
  • iPlant proposal good starting point
  • spring/summer deadlines more realistic
  • convex hulls shapefiles, raster images
  • requirements of what people need for their analyses
  • 4 week lead time to work out sci's problems
  • NSF site visit
  • statistical strength of different models, sampling bias
  • randomization too computationally intense?