UK-SWANSPracticalDigitisation: Difference between revisions

UK-SWANSPracticalDigitisation (view source)

Revision as of 12:37, 22 March 2018

5,185 bytes added , 22 March 2018

→‎Agenda

Dpaul

Bureaucrats

4,713

edits

@@ Line 158: / Line 158: @@
 |
 |}
+== Marine Georeferencing ==
+This section added post-workshop to address outstanding questions in the group about how to georeference (geocode) marine localities.
+=== General guidelines and hints ===
+For a locality like "Puget Anne Sound, 10 fathoms deep", suggestions here from the Georeferencing Working Group (iDigBio GWG) and OBIS.
+#'''Ship and date of collection''' -- if the name of the ship and the date of collection are known, then one can look up the ship logs/itineraries, and find where that particular ship was on that particular day. Sometimes, even coordinates are given in the ship logs. Other times, the ship log will not include specific data for the collecting event, but flanking events will be included, so one can "interpolate" between the two.
+#'''Depth''' -- when just a ballpark area is given, then the depth can help narrow down the uncertainty radius (much like elevation can do this for localities on land).
+#'''Reefs or other ocean floor topography''' are always useful, if they are mentioned.
+#Only other helpful information I can think of are '''bearings given from lighthouses, or even buoys''' (when georeferencing Fish collections from the Great Lakes, it was common for collectors to give these as reference points).
+*Another response to my query shared that marine data is not always so fit for use (better perhaps in OBIS than in other places). So we are looking into how we might help to better reach those using Darwin Core fields for what is expected / needed in particular fields to make these marine localities useful for research.
+*Pieter Provoost p.provoost@UNESCO.ORG, from OBIS, georeferenced "Puget Anne Sound, 10 fathoms deep" and shared his process and results here:
+**We make extensive use of [http://marineregions.org/ MarineRegions], and we also have a simple map tool at [http://iobis.org/maptool/ Maptool] http://iobis.org/maptool/ which connects to the [http://marineregions.org/ MarineRegions] API for georeferencing and makes it easier to produce WKT strings for points, lines and polygons.
+**This is how I would handle "Puget Anne Sound, 10 fathoms deep" for example:
+:locationID: http://marineregions.org/mrgid/15254
+:verbatimLocality: Puget Sound
+:locality: Puget Sound
+:verbatimDepth: 10 fathoms deep
+:minimumDepthInMeters: 18
+:maximumDepthInMeters: 18
+:decimalLongitude: -122.43
+:decimalLatitude: 47.83
+:coordinateUncertaintyInMeters: 20000
+:footprintWKT: POLYGON ((-122.40555 47.56726, -122.50168 47.59505, -122.53601 47.91450, -122.32178 47.90438, -122.40555 47.56726))
+*From Deb: a bit about the above Darwin Core fields. We discussed most of them (briefly!) at the workshop.
+**First, note they have used the locationID field. In the Marine Regions database, certain marine areas have a globally unique identifier. So, using the Marine Regions database, they are populating the locationID field with this unique string. You don't have to use the locationID field, but in situations where the specific region is well defined in Marine Regions it can make good sense to do so - to be very clear about "where" you are. Note that some groups might visit the same marine location repeatedly (right?). And they may have given that marine locality a name, like "Station 67." That's a "local ID" good for communicating a locality to this specific group. IF, "Station 67" provides a unique locationID within the dataset you are publishing, you can put "Station 67" in the dwc:locationID field. In this case, you'll often have local knowledge (maps, journals, ships logs, etc.) about the exact area that is meant by "Station 67" - that you can / will use to georeference that entry.
+**It's clear what Pieter put for verbatimLocality, locality, and verbatimDepth. You can see then that he converted fathoms to meters.
+**Next, he placed a point in the center of Puget Sound and gave it an uncertainty. In essence he put a circle around the point - to enclose all the possible places in Puget Sound - where this specimen could be collected. Note that the value is really a "radius" of the circle (that has as its center, the latitude and longitude provided). The footprintWKT is a set of x,y coordinates that when joined by lines produces a polygon around the lat/lon point. Often polygons allow one to reduce the uncertainty compared to putting a circle around a point. But, they can get quite large (and so difficult to store in a spreadsheet or in a database). The take-home msg for this part is to talk to your IT people.
+Hope you find this useful. Note that Pieter shares they'll add some more examples to the OBIS manual.
+Some more potentially useful bits I found:
+*Short training initiative 2016 http://classroom.oceanteacher.org/pluginfile.php/12805/mod_resource/content/1/Presenting%20the%20trainers.pdf
+*OBIS http://www.iobis.org/ http://classroom.oceanteacher.org/course/view.php?id=319
+*OBIS Simple map tool at http://iobis.org/maptool/
+*Geocoding using the OBIS map tool http://classroom.oceanteacher.org/mod/lesson/view.php?id=7747 (not working yet?)
+*Visualize data points (this is very cool!) http://iobis.github.io/plotter/
+*Georeferencer (overlay old maps with current day!) https://www.georeferencer.com/
 [[Category:Workshop]][[Category: Biodiversity informatics]][[Category:Imaging]][[Category:Project management]][[Category:Data use]][[Category: Collections]][[Category:Digitisation]][[Category:International collaboration]]