2011 working group We Use cases¶

• trait use case on Plone site

Range size¶

• indiv species records from SALVIAS
• min, max lat
• taxonomic cleaning to remove spelling errors
• download from db: species name, lat, long
• precision of coords almost never provided
• WGS 84 coords method
• climate layers w/ 1km resolution
• place name reference already resolved to geocoords
• specimens w/ coords done by georeferencing
• each georeference needs to be recorded
• in DwC, georeference fields
• county occurrence records: political division
• geocoord assoc w/ name
• aggregate vs indiv obs: differentiating
• any db will give name, lat/long
• density-weighted range map
• cultivated specimens: e.g. pines grown in S. America
• metadata relies on whether cultivated
• taxonomic scrubbing
• in VegBank: plot lat/long + stem position (relative or absolute), authorPlantName, maybe taxonInterpretation.plantName
  • stem relative position: needs plot azimuth to resolve to geo coords
  • throw out samples w/o resolved ID
• CTFS plot, lat/long
• calculate lat/long from info: plot coords, plot azimuth, subplot pos, stem pos in plot/subplot
• live with VegBank or use something else?
• one coord for plot + rel coords for stem (more precise than geocoords of stem)
• accommodate w/o changing schema?
• measure coords of plot corners b/c plot can shift over time -> not exactly square
• geocoords pertain to particular plot posts
• some dbs use hexagonal subplots
• subplots lay out the 1 Ha plot
• check if cultivated specimen is column in DwC; if so, add to VegBank: no, it isn't
• need way to flag cultivated vs introduced (DwC fields)
• authorPlotCode: name of subplot
• named corners
• associate lat/long w/ named corner
• protocol to capture way people put stakes in the ground as reference points
• plot has geometry: point, polygon
• VegBank: points are x,y coords relative to plot origin
• add cultivated field
• cultivated vs not is filtering
• import data from source -> filter
• validation pipeline
• model should have DwC fields of specimen desc, locality desc (narrative field; called locality in all versions of DwC)
• VegX superset of DwC
• field notes has info about cultivated: e.g. growing in garden
• specimen desc can be used to detect cultivated field
• DwC leaves out herbarium fields
• species name, lat/long, locale
• cultivated flag goes in taxonObservation table?
• whether cultivated changes if change taxon determination info
• specimen desc not a DwC field
• lat/long: of plot, or of stem?
• centroid of plot w/ error bar
• need lat/long for every stem for some use cases
• VegX currently only manages one ref pt to plot
  • deals w/ verbatim vs derived
• std ways of representing geographic pts, lines, polygons

Measuring species ranges¶

• convex hull, etc.
• input data the same, but different precision requirement
• can't filter by precision b/c data doesn't have precision
• MO data has
• GBIF county data is fuzzy
• geovalidation w/ GNRS (point and polygon validation)
• name resolution to recognize names of political divisions
• query filters by isCultivated, isCultivatedReason, CountyError
• filter by herbarium: tropical trees indoors in Manhattan
• proximity to populated area: too many false positives
• do algorithmic outlier analysis
• lat/long, species name enough? analysis is researcher's task
• placeholder for info that will be computed
  • use annotation to enter that info
• if info in loading script, save it
• location narrative: from DwC locality
• download should have full metadata on ownership to simplify attribution of final data in pub
  • cite each herbarium that contributed data

Disentangling the drivers of beta diversity¶

• species composition in plots
• are morphospecies names consistent between plots
• data needs subplot ID, which isn't in BIEN 2
• BIEN 2 has no spatial nesting capability
• uniform sampling methodology needed
• size cutoff in plot
• VegBank has size bins: stemDiameter, stemDiameterAccuracy
  • lower and upper limits
• how to capture morphospecies scope
• unique code in entire network
• link plots in scope group: fit under project? but scope of morphospecies use not necessarily same for entire project
• VegX doesn't deal with spatial containment
  • uses DwC geospatial distinction?
  • DwC WKT field: well-known text: repr of geometry
• PostGIS has geom column
• convert verbal description to polygon
• sp: undetermined, indet., unknown
• sp for morphospecies? usually unique string w/ site code
• once someone cleans up species name, keep that in db
• specific epithet: need family when genus unknown
• altitudinal gradient: need eleveation
• if data source has elevation, don't rely on DTM data
• locality info usually includes elevation
• elevation approx same across small plot
• need sources
• SRTM (Shuttle Radar Topography Mission) has elevation data for coords (30m accuracy?)
• owner-provided elevation vs. topography elevation

• two more use cases tomorrow

• breaking VegX? no
• changes -> VegX and VegBank
• VegX can deal with subplots
• build db main task; geovalidation is separate task
• need scope of db refactoring
• validation was done to geoscrubbed db
• dealing w/ strata/layers