2012-11-28 breakout groups¶

IUCN¶

• time element
• population change 10-100 yrs
• abundance
• inputs: range size, estimates, land cover/change layers
• range fragmentation
• growth forms
• range size estimates: convex hull, maxent
• decision tree: outcomes -> threatened, vulnerable
• contact IUCN contacts
• global classification of threatened status
• conservation value of species
• species within country boundaries
• workflow
• continental/global scale
• outside feedback on decision tree
• who to best contact
• database of IUCN info
  • transparency: trace back how done
• herbarium specimen data->deduce status
• how many species to detect
• applying redlist criteria to specimens
• range modeling w/ convex hull->range size
• rasterized species distrib
• use ranges as part of tool
• select area->list on the fly
• list, status for any area
• classification of species
• aggregate by vegetation class?

Geo/Eco/Phylo¶

• 10km grid cells
• convert ~40,000 polygon range models to rasters
• aggregate to 100km cells to create easier to work with
• matrix where rows are unique
• presence/absence at 10km resolution, per cell in matrix
• script to unpack matrix to a map
• actual observations
• potential ranges from maxent
• maxent->convex hull->Gentry
• store as compactly as possible: not 12GB R file
• 100s of MB
• probabilistic estimate of presence/absence
• probability estimate
• occurrence data
• need enough precision for that level
• matrix tabulating presence/absence
• matrix interpretation
• shapefile for each species, which follows political boundaries
• raster overlay
• more vs. less inclusive for area
  • most conservative estimate: making sure that you get it if it's there
  • i.e. smaller
  • saying something is rare when it isn't is more problematic
• already is a raster with the sampling intensity John has created
• ultrametracize
• increase coverage in tree: data w/o sequences
• randomly placing species in genera
  • base vs. tip
• matching unplaced taxa
  • non-trivial coding
  • Naim? after 3 mos
• having to do something repetitively
• phylogenetic conservatism of range size
• phylogenetic signals
• range size relationships
• phylogenetic correction
• grid data, plots
• beta diversity
  • descriptive taxonomic/phylogenetic/functional diversity
• what trait maps look like
• community assembly
• relationship between community-weighted traits
• species lists, height data
• remote data
• how tall trees should be
• talk to Jenz
• what species are they? Panama+Peru
• BIEN data, plot data
• signal characters
• trait descriptions
• phylogenetic long-range length
• relate traits to spatial variability and productivity estimates
• cophonetic matrix of tips
• traits, habit data
• conservatism, species measures
• with TRY data
• geographic range, habit
• important part of algorithms
• measures of phylogenetic signals
• algorithm for k
• improved k->accessible framework

DB¶

Validations¶

• candidate datasets
• NYBG data in DwC, look at this afternoon
• plots validations
• refine list
• programming standard part of import
• does extraction conform to what contributed
• data providers are dynamic databases, data changes
• determinations changed->no match
• detect gross discrepancies
• analyze things at given points in time
• outcomes with major economic impacts
• what data working with
• compare data ingested
• compare raw CSV to analytical DB output
• if output = input
• details of validations
• herbarium and plot source
• try everything, the more the better
• 2-part validations
• check that aggregation of individuals is correct
• area in which results will be stored, pointers to source data
• std summarizations, visualizations
• concentrate on papers

4 papers¶

• TREE paper
  • descriptive version
  • popularizing it
• white paper
  • what BIEN is about
  • wider than subgroup
  • more generic than BIEN so would not use BIEN in title
• 1st paper focuses on BIEN 3.0
  • flexible schema for confederating biodiversity occurrences
  • standardization to TDWG standard
• BIEN 4.0
  • what would we do
  • where you would want to go in the future
  • concurrently with BIEN 3.0 paper
  • IEEE/ACM paper: conceptual modeling, DB challenges
  • EER

• next generation of biodiversity occurrence systems
  • build on iPlant proposal
  • pull together data of various types
• generic challenges of where we want to go
• where to go with BIEN
• TREE limitation: authorship number
• bioscience
• concepts changing
• opinion piece in Science or Nature
• paper-related
• feedback to data providers
• what came into BIEN 3
• range checks
  • aspect < 360deg
• 300 data providers
• dbSourceName: verbatim institutionCode
• DwC herbarium codes
• concatenated sources (duplicates)
• critical juncture->send e-mail that data used in project, product is being launched
• refresh data
• send std e-mail, web form show to data provider if don't want to provide data
• send errors to people instead of going someplace to get them
• herbarium acronyms -> contact information
• index by subprovider: include top-level and subsource
• with DwC, data just goes in
• compile from TNRS, geoscrub the list of errors
• add validations

• summaries on Plone site
• draft or outline of project
• WordPress maintain own page
• Redmine site w/ notes
• presentation on TDWG extension in Florence conference
• afternoon subgroups
• habit data
• Barbara's habit scripts
• meet back at 1:30pm

Stephanie on NCEAS¶

• core funding for working groups from NSF
• Packard foundation ecosystems
• topical and core funding
• NSF core funds renewed 5 years ago
• no-cost extension on NSF money
• 11 new working groups next year
• Moore foundation covers core expenses over next 3 years
• conservation, resource mgmt
• press release in Jan/Feb for # new projects
• smaller amounts to fund: scatter of research
• what is on leading edge
• DataONE online
  • big repositories
  • education: summer training workshop
• data discovery and integration
• synthesis
• support from Packard foundation: scholarship model
• summer training
• 4-5 people working on it
• continue beyond 3 years?
• obtain more core NSF funds?
• more working groups, sabbaticals, etc.
• infrastructure for groups with own funding
  • very piecemeal
• convene RCNs
• not a convention center
• collaboration tools to provide to a working group
• suitably themed

Afternoon¶

• 5:15pm Elsies
• 6:30pm group dinner Blue Agave
• whitepaper
• focal points:
  • growth forms
  • ultrametric tree
  • IUCN decision tree
• taxonomy issue
• phenology
• range size issue
• 20 min to validate plots
• validation scripts
• traits
• first groups:
  • IUCN: 2-3
  • range methodology: 3-4
  • validations: rest of afternoon, in main room
  • phenology and traits: next to main room
• later groups:
  • BIEN whitepaper: Bob
• rarity paper
• metadata
• BIEN2 flavors
  • analyses on different flavors

Regroup¶

Traits¶

• phenology paper
• when no info specimen description
• geographic patterns in plant traits
• mean traits, not mean variance
• Plone site has presentation
• papers on herbaria
• parse habit groups
• assess species level phenology for tropical species
• extract tropical info
• ready for comments
• changing functional diversity of plots
• volume niche
• need to write paper in 2 weeks
• functional volume temperate vs tropical
• temperate sites in red, green is tropical, blue is overlap
• functional beta diversity
• mean trait value, overlap of 3D objects
• latitudinal separation
• functional data diversity
• data is from plots

DB¶

• decision tree decisions
• extent of occurrence
• critically endangered categories
• 1 of 6 categories
• use observations in cells
• add pixels to correspond to IUCN criteria
• model grain vs. observations
• occupancy of pixels
• < 5 occurrences
• current, future threat
• clipping model range by land use layer
• for each pixel, desc how far each cell moves and whether acceptible pixel
• under climate change, allow movement into adjacent pixels
• possibilities for local
• not future maxent based on climate change
• assign confidence based on sampling intensity
• attach confidence to category
• start including abundance
• in practice, abundance-weighted measures not used

• project page to WordPress
• use Plone as document and notes repository
• Redmine wiki has running record
• DB enhancements
• code versioning repository
• WordPress site is nascent public interface
  • based on PHP
• technical things in Redmine

Whitepaper¶

• bits and pieces
• test of NY data: one mapping problem, but other than that looks great
• CTFS, Madidi extracts
• many more plots
• will send export of data (different schema)
• target schema
• exchange schema: VegX?
• technological questions
• generic validations
• is pipeline working
• what stage in workflow process where doc takes place
• clear wording
• decide mechanistically how to make this work
• Barbara Dobrin sending scripts to Aaron
  • sent one, adding doct to rest
  • keyword search
  • mining natural language
  • isolated keywords

To do¶

• create workflow to infer taxon habit
• put habit analysis algorithm into BIEN