2011 working group Tu BIEN database¶

• time constraint: 1 year
• what do we want at end of year?
• audience for mapping tool
• additional data not in 2.0 that we want in 3.0
• identify primary science questions to be done w/ BIEN 3
• complexity of plots
• primary users
• educate people on domain
• need MS Access databases for loading data
• plot collection
• present plot data mgmt interface
• kinds of data
  • occurrence data of organism w/ coords
  • plot data
• 50 million herbaria in the world
• digital records of herbarium data
• same species put together
• undescribed species
• metadata: same species in different plots refer to difft things
• specimen is just a presence, but a plot has absences of species, too
• every occurrence/recorded indiv associated with a specific area
• eco plots where only studying a subset of species
• cover if doing all taxa
• safe way to store data in archives
• plot
  • place that doesn't change (geocoordinates)
  • subplots
• plot observations
  • most plots have one obs
  • mult obs if come back to same plot mult times
• lump observations into projects
• community classifications
  • US nat'l veg classification
• taxon obs vs individual obs
• each occurrence of a taxon labeled as a particular kind(s)
• VegBank tables fit into PPT diagram boxes
• VegBank can store as dataset, export to VegBranch
• VegBranch is access tool
• maximize flavors of plots that can go in
  • different vertical strata defined in different ways
  • easy to search, cite, link to, export, import, annotate (errors, community, types of organisms)
• multiple labels: original, current, used in pub
• particular plant might be observed in multiple events
• one name -> multiple specimens, but multiple names/specimen: many-to-many
• three databases
  • plots
  • nomenclature
  • communities
• plug all of VegBank into BIEN 3 or just subset of 7 tables?
• what to add to fit BIEN 3 requirements?
  • traits, other than stem diameter
    • leaf area
    • get from outside b/c linked to taxon name
• plot has locality info
• user-defined attributes
• TROPICOS defines fields: measurements, sla, morphological description
  • attach additional variables
  • traitValue, traitName, traitUnits
• plot properties don't change
  • leave empty unused props
  • location: latitude, accuracy (m or km), confidentiality, public view
• confidential plots b/c on private land: gov't regulate or people take plants
  • fuzzing of location
• cover, strata methods
• metadata will apply to plot level
• embargo rules: period of time or until change status
• real lat vs lat: lat is public view; real is what we know
• coordinate geom: UTM coords converted
• BIEN: might not want to manage private data
• lat/long: origin: corner or center? usually center
  • sometimes multiple for same plot
• reference subplots as pos within plot
• synonyms: mult obs of same thing
  • another name for duplicate
• party boxes: who took data
• get rid of duplicates? but might be slightly different: near synonyms
  • one synonym tagged as current
• how synonyms generated
• some data in CVS database
• additional context to add
• digitization feature like
• data should be identical except for digitization differences
  • e.g. data enterer didn't know that record already in database
• keep inactive records to reestablish link later
• same entity
• cleaned taxa to be consistent with different authority
• observation table has some measurements of place (plot level obs): soil obs, height of tree layer (avg height of canopy), height of shrub layer
• taxon obs are obs of things
• note VegBank diagram legend
• importance values pertain to taxon
• sometimes taxa assigned to vertical strata in taxon observations nest
• stratum is vertical slice of plot by bands of foliage (shrub layer, etc.) or by height
• taxon attributes: cover, biomass
• some sizes done in bins
• taxon observation has name for kind of organism
• count of different-size stems
• record whether tree is leaning; hurricane damage in user-def field
• stem count: # stems in stem size class
• taxon importance
• coord of stem
• CTFS plot just has x, y, height, diameter
• tropical forest plots: taxon name, diameter, height
• but in Carolinas, don't do x, y of every stem: just tally by size class and species
• can do CTFS, cover plots, bins
• lots of column names
• stemCode: tag number
• area overnormalized
• stem x, y position: pertains to one stem or multiple? only for one stem
• VegBank ERD compromise to accommodate multiple protocols; more complicated than necessary
• tree with multiple stems: as same individual: linked together
• strata: merge tables?
• stem counts are aggregated: what is aggregation when aggregate is one
• tuple: species
• people divide between aggregate or individual observations
• indiv vs avg measurements in analysis: throw out obs w/ count >1
• species, # indiv, stem sizes: don't know how indivs map to stems
  • can mix indiv measurements with plot-level measurements

• get VegBank data loading scripts from Steve Dolins
• get MS Access databases for loading data
• get TROPICOS schema
• need desktop dat entry tool for off-the-grid access
  • SQLite, embedded MySQL instance
  • autocomplete taxonomic names
  • modify existing CTFS data entry tool (under development)?
  • talk to Brad
  • Steve has previous implementation
• focus on confederated schema
• modify VegBank based on BIEN 3 analytical db
• check that all repos are there
• takes 4-6 hours to load all data (down from 3 days)
• takes 4 days to load TNRS data b/c querying MOBOT web service 1 name at a time
• Atrium project has online data entry tool