[tdwg-content] Darwin Core: proposed news terms for expressing sample data
Éamonn Ó Tuama [GBIF]
eotuama at gbif.org
Thu Aug 21 10:22:42 CEST 2014
Hi Rob, Anne, Rich,
I think Markus has answered your question as to why we opted for an Event
core which is being used in the sense described by Anne and Rich. For any
event, you can have a list of species in an Occurrence extension and for
each species, you can include quantity and quantityType, e.g., biomass, etc.
The proposed term eventSeriesID was intended for linking together related
events, although it now looks like parentEventID might be a better, more
flexible term. The measurementOrFact extension is a good fit for capturing
environmental information relating to an event. See, e.g., the Gialova
Lagoon brackish water invertebrate test data set [1] where a set of 18
environmental variables, including temp, pH, Rdx, particulate organic
matter, dissolved oxygen, salinity, chlorophyll-a were measured for each
sampling station-sampling period combination. An example mapping is:
Id measurementType measurementValue
measurementUnit measurementRemarks
IA Tmp (sed) 21.5
degree C Tmp (sed): temperature at the bottom
surface
*Controlled vocabularies*
Ideally, the values for samplingUnit and quantityType would be selected from
controlled vocabularies. This is, effectively, what we do by presenting a
small list of values in a drop-down menu. The current values are what we
derived for example data sets and discussion but they can undoubtedly be
extended and improved.
We capture “bucket” type measures through a combination of samplingEffort,
samplingGeometry and samplingUnit. For example, a pitfall trap (in a point
location) left out for 16 days might have samplingEffort: 16,
samplingGeometry: point and samplingUnit: day. Three m^2 quadrats in a shore
survey might have samplingEffort: 3, samplingGeometry: area and
samplingUnit: m^2.
It would be very useful to see your compilation of scope, effort and
completeness measures to see if we can express them in our model and/or if
we need to reconsider our approach.
Éamonn
[1] http://eubon-ipt.gbif.org/resource.do?r=ionian-brackish-lagoon
From: tdwg-content-bounces at lists.tdwg.org
[mailto:tdwg-content-bounces at lists.tdwg.org] On Behalf Of Markus Döring
Sent: 20 August 2014 23:47
To: Robert Guralnick
Cc: TDWG Content Mailing List
Subject: Re: [tdwg-content] Darwin Core: proposed news terms for expressing
sample data
Rob,
this proposal if for monitoring surveys really, not to be confused with
material samples like environmental or tissue samples which have a distinct
new dwc class MaterialSample.
We tend to overload the term sampling a lot and it helps treating material
samples different from pure observational "sampling". That is why the
existing Event class was used as the core and classic Occurrence records as
extensions. A classic example is a vegetation survey where each plot
represents an Event record and each recorded species in that plot will be an
Occurrence extension record with a given quantity. Darwin Core already
offers individualCount to specify quantity, but it is a very specific way of
measuring "abundance" restricted to only some use cases. Abiotic
measurements about the plot (e.g. soil type, pH, temperature) can be
published using the measurements or facts extension linked to the Event
core.
Markus
On 20 Aug 2014, at 20:08, Robert Guralnick <Robert.Guralnick at colorado.edu>
wrote:
Anne -- I don't know the answers! These are questions for Eamonn. I
would presume that a sample could be a jumble of species or even just water
or soil samples, and biomass would refer to that sample - but maybe that
isn't a use case being considered? The examples given in the longer
document all link an event_id to species name and some measure of quantity
for that species (to the species, not an individual specimen), so I assume
that is the prevailing (or only) case?
Best, Rob
On Wed, Aug 20, 2014 at 11:56 AM, Anne Thessen <annethessen at gmail.com>
wrote:
Hi Rob
I would like to respond to your item number 2.
>From my perspective, I deal with lots of published descriptions of taxa. The
text might say something like "I saw species A in the Chesapeake Bay, the
Adriatic Sea and the Indian Ocean and the biomass is 5 - 9 grams". The
biomass range obviously corresponds to at least three different occurrences,
but how to divide the biomass data? I would love to be able to have an
*event* to attach it all to. There is almost two different levels of events
- a sampling event and a "study event". The "study event" would correspond
to the type of event I would like to use in the above example. It may not be
ideal, but for the old literature that might be the best we can do.
I have to admit that I don't know enough about trawl data to understand why
an event core would be a problem. It seems that the trawl would be an event
and each biomass measure (of each fish) would be attached to a separate
occurrence which is attached to that event. Am I understanding this wrong?
btw - I found a workaround for the example I gave, so it's not impossible to
model with the current structure....
Anne
On 8/20/2014 1:16 PM, Robert Guralnick wrote:
Éamonn et al. --- Thanks for the clarifications. I think these help a ton
but it raises a couple more questions for me.
1) I am surprised that you plan to use of MeasurementorFact extension in
relation to the Event core, which seems like a novel (or perhaps awkward or
unintended?) mechanism for capturing environmental data, but the same
extension was not be seen as relevant for describing samples? Can you
explain more about the thinking there?
2) There may be a subtle issue here extending "Event" to be more what you
call a "Sampling Event Core". My read of this is that Darwin Core serves as
a way to deal with point occurrences and Event reflects the context of a
single capture event (whether a single observation, or a bulk sample
capture). The changes recommended seem to dramatically extend and change
that meaning? Its simply a question that I don't have answer to, but is
Darwin Core, the right vehicle to start capturing repeated measures of
biomass values from trawls? I don't have answer but man, terms like
quantityType (as a property of occurrence?) give me pause.
3) Is Sampling Unit a controlled vocabulary? For another project, I have
looked through - and captured scope, effort and completeness measures from -
a large number of published biotic area inventories. The vast majorities of
these are measured in units like bucket hours, or trap nights. Is a
"bucket" part of SamplingGeometry or Sampling Unit? I'd be happy to send
along all the many examples of how biotic inventories of an area are
completed and perhaps it might be good to see how those might be represented
using the terms you are proposing?
Best, Rob
On Wed, Aug 20, 2014 at 10:16 AM, Richard Pyle <deepreef at bishopmuseum.org>
wrote:
Same here – Events are central to the work that we do.
Aloha,
Rich
From: tdwg-content-bounces at lists.tdwg.org
[mailto:tdwg-content-bounces at lists.tdwg.org] On Behalf Of Anne Thessen
Sent: Wednesday, August 20, 2014 2:59 AM
To: tdwg-content at lists.tdwg.org
Subject: Re: [tdwg-content] Darwin Core: proposed news terms for expressing
sample data
Hello
I would just like to comment on *event core*.
I've been doing a lot of work translating published data into Darwin Core.
During that process I've wished several times that I could use Event as
core. I am happy to hear about that proposed change. It will make it easier
to model the data I am working with.
Anne
On 8/20/2014 7:04 AM, Éamonn Ó Tuama [GBIF] wrote:
Hi Rob,
Thank you for the feedback. I have tried to address the two main issues you
raise below. At the outset, I would like to emphasise that much of this work
is taking place in the context of the EU BON project which includes a task
on developing/enhancing tools and standards for data sharing with a
particular focus on the IPT for publishing sample-based data. So, we were
constrained by the need to publish sample-based data sets in the Darwin Core
Archive format and to demonstrate practical application using a working
prototype. When the discussion on the TDWG list faded out, we took it to our
EU BON partners whose requirements were essential input to further
development. We recognise that these discussions took place away from TDWG
(although the TDWG/EU BON contributors overlapped) and this is the reason we
are presenting the outcomes here for further consideration.
*Event core*
As the SIGS report indicated, sample data can be modelled in Darwin Core
Archives using either Occurrence or Event as core. This was the starting
point for our evaluation but as things progressed the data wrangling pushed
the model back towards the Event core. We actually went through the exercise
of mapping multiple test datasets in an iterative process spanning several
months' work. In the end, we found that using an Event core better matched
the typical sample data we were dealing with, allowing use of a
measurement-or-fact extension to be included for the efficient expression of
environmental information associated with the event. The choice comes down
to an Occurrence core or an Event core + Occurrence extension. In both
cases, the true observation records are Occurrences. The big difference is
what type the core has and therefore to which kind of records you can attach
further facts and extra information with DwC-A extensions. Many sampling
datasets have very rich information about the site and event, so it is very
natural to hang facts from an Event core. When picking the Occurrence core
those facts would have to be repeated for each and every occurrence record.
Moreover, our approach doesn’t stop anyone from using the Occurrence core if
they so wish. This just provides a different option for datasets that better
fit an Event core model.
I want to stress that we are not building a “specific IPT version” to
support an Event core but, rather, we adapted the IPT so that it can be
configured to support any generic “core + extension” format to enable its
use for exploration of more data formats. This is part of the core codebase
and there were no custom forks of the IPT for this work. Our view at GBIF
is that if there are significant numbers of data publishers who are keen to
adopt, promote and use a (any) format, and the tools can be configured to do
so, then we should support it, and, if necessary, use a custom namespace.
*New terms around abundance*
Yes, the discussion on TDWG did fade out but it was clear that the term
“abundance” as recommended by the SIGS report (along with
abundanceAsPercent) was confusing many when we were looking for term(s) that
reported quantitative measures of organisms in a sample. It also became
clear we would need to be able to state the type of quantity being measured.
An alternative suggestion for using the MeasurementsOrFact class was
immediately shot down.
As some of our main use cases were coming from the EU BON project,
discussion shifted to that forum and consensus formed about the currently
proposed terms. It was within this group that the additional terms
(samplingGeometry, samplingUnit, eventSeriesID) were proposed and where we
began testing with sample data sets.
Best regards,
Éamonn
From: robgur at gmail.com [mailto:robgur at gmail.com] On Behalf Of Robert
Guralnick
Sent: 19 August 2014 16:56
To: Éamonn Ó Tuama [GBIF]
Cc: TDWG Content Mailing List
Subject: Re: [tdwg-content] Darwin Core: proposed news terms for expressing
sample data
Hi Éamonn --- I am curious about the outcomes presented in the SIGS paper,
in particular, this portion of the paper:
"Solutions without introducing an event core in Darwin Core Archives:
During the review of the solutions for the uses cases, it became apparent
that either model could be applied to every use case. The core and
extensions bore a complementary relationship and between them could express
all the required information. The core simply provided the central anchor in
the star schema from which to join the additional information. Therefore,
using the Occurrence core, well established in the GBIF network through
uptake of the IPT, seemed more appropriate than inventing CollectingEvent as
an additional core type."
That SIGS paper has John Wieczorek and you both as authors, including
many luminaries across the biodiversity standards spectrum. Given the
above, its curious to see the EventCore come back again, along with a
specific IPT version to support it.
So I see two issues, conflated, in this post you just made. One is the
need for an EventCore at all, and the nature of relating Event and
Occurrence/Material Sample. The second is the introduction of new terms,
which seemingly have arrived after debate on similar terms - but framed
around abundance - stalled a year ago. To my mind, these both require some
further discussion, because I don't (necessarily) see TDWG community
coherence around either issue?
Best, Rob
On Tue, Aug 19, 2014 at 6:11 AM, Éamonn Ó Tuama [GBIF] <eotuama at gbif.org>
wrote:
Dear All,
GBIF is committed to exploring ways in which the IPT and Darwin Core Archive
format can be extended for publishing sample-based data sets. In association
with the EU BON project [1], a customised version of the IPT [2] has been
deployed to test this using a special type of Darwin Core Archive in which
the core is an “Event” with associated taxon occurrences in an “Occurrence”
extension.
The Darwin Core vocabulary already provides a rich set of terms with many
relevant for describing sample-based data. Synthesising several sources of
input (GBIF organised workshop on sample data, May 2013 [3], discussions on
the TDWG mailing list in late 2013; internal discussion among EU BON project
partners), five new terms relating to sample data were identified as
essential. The complete model including these new terms are fully described
with examples in the online document “Publishing sample data using the GBIF
IPT” [4].
As a first step towards ratification, we would like to register the new
terms in the DwC Google Code tracker [5] if there are no major objections on
this list. The five terms are:
1. quantity: the number or enumeration value of the quantityType (e.g.,
individuals, biomass, biovolume, BraunBlanquetScale) per samplingUnit or a
percentage measure recorded for the sample.
2. quantityType: : the entity being referred to by quantity, e.g.,
individuals, biomass, %species, scale type.
3. samplingGeometry: an indication of what kind of space was sampled;
select from point, line, area or volume.
4. samplingUnit: the unit of measurement used for reporting the
quantity in the sample, e.g., minute, hour, day, metre, metre^2, metre^3.
It is combined with quantity and quantityType to provide the complete
measurement, e.g., 9 individuals per day, 4 biomass-gm per metre^2.
5. eventSeriesID: an identifier for a set of events that are associated
in some way, e.g., a monitoring series; may be a global unique identifier or
an identifier specific to the series.
Best regards,
Éamonn
[1] http://eubon.eu <http://eubon.eu/>
[2] http://eubon-ipt.gbif.org <http://eubon-ipt.gbif.org/>
[3]
http://www.standardsingenomics.org/index.php/sigen/article/view/sigs.4898640
[4] <http://links.gbif.org/sample_data_model>
http://links.gbif.org/sample_data_model
[5] https://code.google.com/p/darwincore/issues/list
____________________________________________________
Éamonn Ó Tuama, M.Sc., Ph.D. (eotuama at gbif.org),
Senior Programme Officer for Interoperability,
Global Biodiversity Information Facility Secretariat,
Universitetsparken 15, DK-2100, Copenhagen Ø, DENMARK
Phone: +45 3532 1494 <tel:%2B45%203532%201494> ; Fax: +45 3532 1480
<tel:%2B45%203532%201480>
_______________________________________________
tdwg-content mailing list
tdwg-content at lists.tdwg.org
http://lists.tdwg.org/mailman/listinfo/tdwg-content
_______________________________________________
tdwg-content mailing list
tdwg-content at lists.tdwg.org
http://lists.tdwg.org/mailman/listinfo/tdwg-content
--
Anne E. Thessen, Ph.D.
The Data Detektiv, Owner and Founder
Ronin Institute, Research Scholar
443.225.9185
_______________________________________________
tdwg-content mailing list
tdwg-content at lists.tdwg.org
http://lists.tdwg.org/mailman/listinfo/tdwg-content
--
Anne E. Thessen, Ph.D.
The Data Detektiv, Owner and Founder
Ronin Institute, Research Scholar
443.225.9185
_______________________________________________
tdwg-content mailing list
tdwg-content at lists.tdwg.org
http://lists.tdwg.org/mailman/listinfo/tdwg-content
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.tdwg.org/pipermail/tdwg-content/attachments/20140821/1e1f1db0/attachment.html
More information about the tdwg-content
mailing list