18: Image data model Resolution:
Discussion: Look for existing protocol in the future to potentially use.
19: Expertise data model Resolution:
Discussion: Look for existing protocol in the future to potentially use.
21: How sure of taxonomic id? Resolution:
Discussion: - TaxonomicConfidence: Sure, Guess?
- Accuracy is given by level of taxonomic id: species, genus, subspecies, variety
- needs investigation
- Talk Bob Peet
64: Have a Metadata DataModel?. Resolution:
Discussion: Jim: Does TDWG have something we can use here? The TAPIR MetaData? DataModel? is for the entire data source so it does not fit well.
This may partly be covered by existing data models. Some providers, though, get their data from a network of providers. Example of missed information was a website with a database of projects.
82: Where does look-a-like data fit into the protocol? Resolution:
Discussion: Providers have data on look-a-like species(e.g., this species is like this invader so maybe we should look out for it). Similar species - future invader or non-invader? Where do these data fit into the protocol?
52: Add a new data model for EnvironmentalInfo? Resolution:
Discussion: This datamodel will help distinguish between species that occur in natural or human modified environments, and/or in freshwater, brackish, marine and terrestrial habitats. Multiple values are possible. These concepts are different from EcosystemImpacted? and HabitatImpacted? (a terrestrial species can impact aquatic environments). At the GISIN portal someone could filter by ‘terrestrial‘ to get all terrestrial species – whether or not the habitat impacted might be both terrestrial and aquatic.
EnvironmentalInfo supports the following additional Concepts: Possible values for the additional Concepts are listed in the table.
67: How will we represent continents and oceans? Resolution: We will add ContinentCode and will not add dac:Continent
Discussion: Not going to add ContinentCode at this time, waiting on working group.
We got a set from Olivier - UNLOCCODE; Look at FAO
Formed a working group inlcuding Annie, Olivier, and Christine.
85: Database of documents Resolution: Add a resource type of 'Other document'
Discussion: Where do these data fit in the protocol? These data are different from the Species Resouce URL Resource Type of reference. These would be actual documents rather than a citation.
90: Need suggested text fields for ImpactStatus Mechanism concept Resolution: See new Mechanism concept in ImpactStatus.
Discussion: Michael provided a list of negative impacts: Allergenic, Competition, DiseaseTransmission?, Herbivory, Hybridisation, InteractionWithOtherInvasiveSpecies?, Pathogenic/Parasite, PhysicalDisturbance?, Predation, SoilTransformation?, Transpiration, Other, Unknown
2: How to maintain source/credit/authorship/stewardship/originator? Resolution: Add RecordID to each record, PreviousRecordID when needed, and add Citation DataModel. Already have
Discussion: Add RecordID to each record, PreviousRecordID when needed, and add Citation DataModel. Already have Citation text field.
Discussion: Why: Acknowledge everyone in the chain. The citation text field should contain an appropriate citation from the original provider. Providers who change the data can add to this field but must maintain it’s original content. Each record can have a RecordID that is registered with an LSID authority. The information at the authority will contain information on the provider. A provider who changes data from another provider must retain the previous providers record LSID as “PreviousRecordID”. This allows the chain of providers to be examined. The citatation data model will be defined in the future for more complex, searchable citation information. Providers relaying information from other providers must maintain the Citation and PreviousRecordID fields, otherwise the field defined here are recommended but not required. The citation text field should contain “How I could like it to be cited” from the providers.
3: Do we need to represent intermediaries? Resolution: See issue 2
Discussion: - Modelers will want to know the whole chain?
- Receive changes and corrections?
- “credit string”
- “Do you want to be cited?”
- “How do you want to be cited”
9: There are many other spatial regions at least for BioStatus/DistributionRecord. Which do we encode? Resolution: See discussion
Discussion: 1) We will add LocalityType, LocationStandard, LocationValue
- LocalityType: defines the type of info in the Locality concept (city, park, cemetery, refuge, protected area)
- LocationStandard: defines the standard used by the provider to identify the location of the record (i.e. US_HUC, US_FIPS, AR_PostalCode, etc.). In GISIN 1 we only support one standard location codes per record. The provider must select the most appropriate one. Providers may provide one LocatoinStandard/LocationValue pair and any number of other location concepts.
- LocationValue: 8001, etc.
2) We will add International and National standards for location codes as available and appropriate and will maintain a list of recommended codes.
3) We will add WaterBody, IslandGroup, and Island from DC and encourage the use of Standard Terms for these concepts.
4) We will use international and national standards where available. We will create international standards where needed. Location standards will be placed in an appendix and will include:
- US_FIPSCode
- US_HUC
- AR_PostalCode
- GISIN_IslandGroups
- GISIN_Islands
- GISIN_FreshWaterBodies
5) LanguageCode is required for any non-standard concepts (i.e. Locality, WaterBody, etc.)
Discussion:
- “Island group”
- “Native to south America”
- “Native to Asia”
- “Eastern Caribbean”
- “Tropical Central America”
- “Himalayas”
13: Add a “Title”, “Description”, and “URL” for each record? Resolution: Just a Title and Description to SpeciesResourceURL
Discussion: Memo is used for comments, notes, information about the record and data that does not fit into the other defined concepts. Description is about the content of the resource and is appropriate to display with the title.
20: Need way of indicating derived occurrence coordinates? Resolution: Use the GeoreferenceProtocol concept in the GeoSpatial Extension but with a controlled vocab
Discussion: The GeoreferenceProtocol is not a controlled vocabulary.
- “Reference to the methods used to determine the coordinates and uncertainties”
We will define our own controlled vocabulary and recommend it to the DarwinCore group.
Defined Types:
- GPS
- Map (GoogleMaps with Map, MapQuest, Topographic Maps, Street Map)
- Gazeteer (GoogleMaps search)
- ConvertedFromOtherUnit
- RemotelySensed
- GIS derived (e.g., centroid calculation for a state/ province)
28: Can toolkit do the translation to UTF8 w/extensions? Resolution: The toolkit can do this translation as it is relatively simple and
Discussion: PHP can translate from Unicode 16 and 32 to UTF-8 (ASCII) and does provide HTML/XML style enoding. This should allow the toolkit to translate data from a Unicode (multibyte) database entry into an ASCII stream with HTML/XML encoded escape sequences.
29: Which UTF8 extension standard will we use? Resolution: Use HTML and XML standard escape squences such as XXX; (hex) and DDDD; (decimal)
Discussion: - Need to find out what the standard is for encoding UTF16 in UTF8
- Use HTML extended for now
- The XXXX; (hex) and DDDD; (decimal) are the standard for HTML and XML to encode Unicode 16-bit and 32-bit characters into an 8-bit character scheme.
37: Add temporal precision? Resolution: TemporalPrecision: Year, Month, Day
Discussion: Which level of date: day, month, year
39: How do we support SOAP requests? Resolution:
Discussion: Olivier volunteered to investigate a SOAP portal that will allow folks to call a server and the server will call the providers and return SOAP data
43: Will BioStatus and Occurrences remain separate data models or be combined? Resolution: Current solution is to have them be separate
Discussion: Proposal: They will remain separate but BioStatus will be renamed “SpeciesStatus”. Occurrences are a location, taxon ID, and date for an occurrence for a species. Sometimes occurrences will also have BioStatus and in this case they will appear as BioStatus as well. The main reasons for maintaining the separation is to 1) keep from having to respond with all occurrences when a consumer only requests BioStatus and to 2) maintain some compatibility with TDWG/GBIF for occurrences.
- Annie says they should remain separate.
- Separate makes occurrences closer to DiGIR
- Default would be to get status with all occurrences
- Default would be to get all occurrences when you request biostatus
51: Should we add Resolution: Yes but internally managed for toolkits, not part of protocol
Discussion:
53: Change Harmful back to Invasive in BioStatus? Resolution: Nope
Discussion: Discussion: Many believe that ‘Harm’ is subjective and that it is the culmination of the invasion process, which begins with establishment, then persistence and spread. If we provided the word ‘Invasive’ instead of ‘Harmful’ in the BioStatus? concept, most providers would be able to map to its possible values, but each provider might mean something different. The key differences are between;
ONE: Scientific uses of the word ‘Invasive’ meaning ability to spread (i.e. Distribution = Widespread or Moderate, and/or Abundance= Dominant or Common) and
TWO: Policy uses of the word ‘Invasive’ meaning ability to cause harm.
Different meanings or usages can be deduced from the additional concepts for which data is provided. Currently, the first group can select ‘Invasive’ then map their data to Distribution = (Widespread or Moderate), and/or Abundance= (Dominant or Common) and/or Harmful= (Yes or Potentially).
The second group can select ‘Invasive’ then map their data to ImpactStatus? concepts as well as those of group 1.
More Discussion: Folks had major issues with the term “Invasive”. We decided to use “Harmful=Yes”. However, invasiveness data would not be mapped to the ‘Harmful’ concept if there is only evidence of establishment, persistence and spread (e.g. when harm has not been confirmed). A search at the GISIN portal on Harmful = yes would miss this invasiveness data.
55: Introduce an ImpactMechanism? concept Resolution: Refer to Issue #88
Discussion: Introduce an ImpactMechanism? concept to handle information about competition, predation, etc. i.e. how IAS damage species habitats and ecosystems. This kind of information is widely available and ImpactMechanism? allows us to make assumptions about which kinds of native species may be threatened.
58: Introduce a DispersalStatus? Concept called MovementStatus? Resolution: Added.
Discussion: DispersalStatus? applies to species movements at all scales – introduction across international borders as well as dispersal from one watershed to the next. We could introduce a DispersalStatus? Concept called MovementStatus? with possible values = Pre-borderMovement and Post-borderMovement.
59: Introduce Concepts for DispersalMechanisms? and DisperslPathways? Resolution: Removed cause and vector, modified other concepts - refer to issue #87
Discussion: The problem with the original Cause and Vector concepts is that it would have been difficult for providers to map to values as there are so many. Now they can map to 1 of 3 dispersal mechanisms and/or 1 of 6 dispersl pathways. This simplified approach lends itself to comparative analysis across a wide range of taxa and to policy applications (see the ‘responsibility’ statement in the Descriptions for Pathway values).
62: What format should we use for dates? Resolution: Use the DarwinCore standard date format (ISO 8601 date and time)
Discussion: Discussion: Michael: YYYY-MM-DD is usually unavailable for ‘Date of introduction’. Typically providers have a year or a decade. How do we handle the lack of MM-DD? Discussion: Michael: You often get pre- or post- a year or a decade. Could we implement something modelled on DAISIE, which uses 2 fields: If the date is precise, the same date appears in both fields. If the first field alone is populated, the meaning = ‘post the date’, if only the second field is populated, the meaning = ‘pre the date’.
Discussion: Jim:We can setup the toolkit to map years (including a decade) to a date field. Mapping multiple columns into a date is more complicated. All databases have standard date fields that can be automatically mapped to the protocol. On this one I think I would suggest we recommend the providers use the SQL standard date fields if at all possible.
69: What is the minimum information required to display a record that uses a location reference? Resolution: GISIN will accept anything provided by the provider.
Discussion: We will ask providers to populate as much information as they can within the protocol (i.e. we only have fields for one standard location value). The portal may filter out incomplete hierarchies. Either a Lat/Lon and Datum or Geographic hierarchy is strongly recommended. Examples of Geographic Hierarchies include:
- CountryCode
- CountryCode, State
- Country,Code State, County
- CountryCode, State, County, LocalityName [city, river, road, park, etc.]
- ContinentCode
- ContinentCode, LocalityName
Note: Any of the valid hierarchies can end in a LocalityName
72: Which country codes should we use Resolution: ISO 3166-1 Alpha-2 codes - will match Darwin Core
Discussion: 2 letter version is the most widely used. Fr-ca for French Canadian needs to be investigated. Three letter codes have less chance of overlapping and are clearer in some cases.
79: Add Virus as a kingdom Resolution: Add 'Other' as a kingdom option because kingdom should be a required field.
Discussion: Viruses are not represented in standard kindoms. Proposal is to add viruses as a kingdom. Or add an 'Other' field with description including viruses, bacteria, and other organisms not inlcuded in current kingdom structure.
86: Changes to ManagementStatus Resolution:
Discussion: Add a value to action of 'None'
Add a new concept of 'Outcome' with controlled vocabulary of Successful, failed, and unknown.
Outcome should be unknown unless status=completed.
We will check on common terminology and may change the controlled vocabulary to match more commonly used terms.
87: Changes to DispersalStatus Resolution:
Discussion: 1) Maintain Route - define as descriptive sentence, e.g., this species was brought over on a coal ship...
2) Rename concept Method to Mode following Michael's suggestion with potential values of Natural, Deliberate, Accidental, Unknown
3) Create new concept Movement following Michael's suggestion with potential values of IntraRegionalMovement (pre-establishment), InterRegionalMovement (post-establishment), Unknown
4) Create new concept of IntroducedBy with potential values of Individual, Government (would include most military, state universities), PrivateSector (for profit; would include private academia like Elmira College), InternationalOrganization, Other (this would inlcude NGO, pirate, church), Unknown
5) FromCountryCode - keep that, but also add other 'To' common location fields with 'From' in front; missing 'LocalityType', 'LocationStandard', and 'LocationValue'
6) Add a concept 'Mechanism' that would follow the Impact::Mechanism where we have a list of suggested terminology to use: Ballast water, hull fouling, aquarium trade, aquaculture, intentional release (stocking), intentional release (stocking), escape (horticulture), agriculture, pet trade, hitchhiker, natural (wind, water), soil movement, floating vegetation, canals, internet sales, roads, railroads
Working group will revise Mechanism list - Brian, Christine, Chris, and Michael.
7) Remove current Cause and Vector concepts.
88: Changes to ImpactStatus Resolution:
Discussion: Change HabitatType to Biome.
Change concept Ecosystem to Environment, add a third value Artificial.
Change concept HarmType to Target - can have some benefit.
Change value types to be Environment, HumanHealth, Socioeconomic.
Change ImpactStrength to Strength.
Add new concept of Direction with values of Neutral, Positive, and Negative.
Add new concept Mechanism with un-controlled vocabulary (string field where multiple mechanisms can be added) but a list of suggested vocabulary to consider.
91: GUIDs for records Resolution:
Discussion: • Will strongly recommend persistent, resolvable GUIDs
• Will accept existing GUIDs, if provided
• If not provided, will recommend DwC-recommended standard and authority for persistent and resolvable GUIDs, and make it easy to adopt them. GBIF or GISIN will keep a registry of institution codes. www.gisin.org/guid will be the authority for now.
GUID permalink-
gisin.org/guid/[InstitutionCode]/[CollectionCode]/[CatalogNumber]
Example: gisin.org/guid/niiss/godm/12
ACTION: Kyle will ask if GBIF is keeping a catalog of InstitutionCodes to use as an authority.