Difference between revisions of "Category:Trees Working Group"

From Linked Earth Wiki
Jump to: navigation, search
( Pages with a poll, Working Group )
m (fix typo)
m
 
(22 intermediate revisions by 3 users not shown)
Line 19: Line 19:
 
* provide separate recommendations for new and legacy datasets
 
* provide separate recommendations for new and legacy datasets
  
 +
== Scope ==
 +
 +
Tree rings are used by many different researchers, for many different reasons.  The LinkedEarth community is well rooted in the paleoclimate community and while the ontology and data standard has great potential for use outside this community, we need to decide how to address the overlap in our discussions.  Should be concentrate on the needs of the paleoclimate community exclusively to begin with, and intend to go back and revise at a later date?  Or should we attempt to accommodate all potential users of tree-ring data from day 1?
 +
 +
The discussion within this group can build upon the work done by the Tree Ring Data Standard (TRiDaS) consortium <ref>Jansma, E., Brewer, P. W. & Zandhuis, I. (2010) TRiDaS 1.1: The tree ring data standard. Dendrochronologia 28, 99-130. [http://dx.doi.org/10.1016/j.dendro.2009.06.009 DOI]</ref>.  While the contributors were weighted towards dendroarchaeologists, the standard was intended to be universal for all uses of tree-ring data.  The requirements laid out by the TRiDaS contributors are therefore going to largely be relevant to discussions here.
 +
 +
'''Note''': Do not edit the polls (even for typos) once voting has started as it will reset the vote counts to zero. If a change needs to be made, make an annotation above the old poll (i.e., above the poll tags) and place the new poll below the first one.
 +
 +
<poll>
 +
What should the scope of this working group be?
 +
Solely the needs of the paleoclimate community
 +
Concentrate on the needs of the paleoclimate community initially while keeping in mind wider needs
 +
Handle the needs of as many communities as possible
 +
</poll>
  
 
== Sensors ==
 
== Sensors ==
 
An initial recommendation is to focus on different sensors:
 
An initial recommendation is to focus on different sensors:
 +
 +
=== Measurement variables ===
 +
 +
The traditional variable measured in dendrochronology is of course the whole ring-width.  It is common now to collect a wealth of other variables from tree-rings including:
 +
 +
* Whole ring width
 +
* Early wood width
 +
* Late wood width
 +
* Whole ring density
 +
* Early wood density
 +
* Late wood density
 +
* Maximum density
 +
* Latewood percentage
 +
* Vessel size
 +
* Blue intensity (density proxy)
 +
* Stable isotopes
 +
 +
=== Event information ===
 +
 +
Tree-rings provide the potential to store information about specific events.  The most common of these is a forest fire (stored as a fire scar or other anatomical damage), but may include the effects of defoliating insects, volcanic eruptions etc.  This sort of event data is represented as binary yes/no occurrences in particular rings and is quite distinct from the continuous measurement variables (e.g. ring-widths, isotope concentrations etc) that are typical with most proxies.
 +
  
 
=== Tree rings ===
 
=== Tree rings ===
 +
  
 
=== Detrending ===
 
=== Detrending ===
  
How to document the type of detrending applied to enable reproducibility.
+
Maintaining references back to raw ring-width measurements is important to maintain data transparency.  When detrending has been performed a detailed record of what has been done should be stored along with the software/library used.  Implementations of different detrending techniques in different software can produce different results. To ensure results can be replicated the following fields will be necessary in addition to the detrended data:
  
=== Stable water isotopes ===
+
* Software used
 +
* Version used
 +
* Detrend method name:
 +
** Polynomial
 +
** Exponential
 +
** Moving average / Floating 
 +
** High pass filter
 +
** Cubic spline
 +
* In the case of some methods (e.g. polynomial, moving average) then a parameter may be required.  I can only think of methods that require one parameter, but for future proofing it would make sense to be able to pass multiple parameters.
  
 +
=== Stable isotopes ===
  
Here are polls that the group might want to consider:
+
Some recommendations were made for an 'International Tree Ring Isotope Databank' ITRIDB by Csank<ref>Csank, A.Z. (2009) An International Tree-Ring Isotope Data Bank– A Proposed Repository for Tree-Ring Isotopic Data. Tree-Ring Research 65(2):163-164. [http://dx.doi.org/10.3959/1536-1098-65.2.163 DOI]</ref>
  
 +
With tree-ring isotopes workers recognized that in addition to the information provided with a ring width dataset there is a need to provide several extra pieces of information as metadata . In addition the usual information on site name, investigator, publication information, genus/species, latitude/longitude, etc. the information that was deemed important to include with any tree-ring isotope submission is listed below.
  
<poll>
+
* Whether or not a chronology was established
What is your favorite poll?
+
 
A poll that is droll
+
* If there is no chronology (or it is floating in the case of sub fossil material) how was age established.
A poll with a role
+
 
A poll with a roll
+
* If it is tropical tree data how was the age model established and what are the uncertainties associated with that age model.
</poll>
+
 
 +
* Were samples pooled (either single year/single tree) or are samples single ring/single tree measurements that have been averaged
 +
 
 +
* What portion of the ring was analyzed (e.g. was it whole ring, earlywood, latewood middle ring)
 +
 
 +
* Were were samples pre-treated by conversion to alpha cellulose, holo cellulose or in the case of <sup>2</sup>H nitrated cellulose or was unprocessed wood analyzed
 +
 
 +
* If pre-treated, what method of pre-treatment was used (e.g. Leavitt and Danzer, 1993; Loader et al. 1997;  Brendel et al. 2000; modified Brendel (Anchukaitis et al. 2008); Rinne et 2005; Li et al. 2011; Weiloch et al. 2011; Kagawa et al. 2015)
 +
 
 +
* What was the analysis date
 +
 
 +
* Which isotope(s) was(were) measured (e.g. <sup>2</sup>H, <sup>13</sup>C, <sup>18</sup>O, <sup>15</sup>N, <sup>34</sup>S)
 +
 
 +
* Analytical precision for each isotope analyzed
 +
 
 +
* If ẟ<sup>13</sup>C data is presented was it corrected for the Suess effect or not and if so which correction was used (e.g. McCarrol et al. 2009) (''one suggestion is that only uncorrected data should be presented and users can correct their own data as is done for <sup>14</sup>C data'')
 +
 
 +
* Which stable isotope laboratory performed the analyses.
 +
 
 +
== Location information ==
 +
 
 +
Prior to the wide availability and affordability of GPS handsets, location information in dendrochronology tended to be low accuracy coordinates for a particular study site.  This is what is stored in the ITRDB.  However, for many years now the many (majority? all?) researchers have collected GPS coordinates for each individual tree.  A clear distinction is needed in the ontology to record what the location information is recording.
 +
 
 +
Chronologies are by definition an amalgamation of data from multiple trees.  Location information for these can therefore be represented either as multiple coordinates (one for each tree) and/or a polygon denoting the extent of the area covered by the chronology.  Should such a polygon be the smallest area encompassing all tree coordinates or should it be the freehand area covering the area the researcher feels the chronology represents?
 +
 
 +
Further complications arise when handling tree-ring data collected from non-living trees.  Sample from snags can be reasonably expected to be at least close to where the tree grew, but archaeological and other cultural samples can be found miles from their initial growth location.  Very long range transport/trade of wood has taken place for millenia.  Some studies have successfully determined the provenance of transported/traded samples through dendrochronological analyses and/or isotopic work.  In these cases it would be good to also be able to store this sort of information.  Regardless though, the type of location must be clearly stored with all locality data.
 +
 
 +
 
 +
== Taxonomic information ==
 +
 
 +
Taxonomic information is clearly essential for tree ring (and other biological proxies).  The current most comprehensive database of authoritative taxonomic information is the [http://catalogueoflife.org Catalogue of Life].  As of 2016 it contains over 1.6m species from 158 contributing databases covering plants, animals, fungi and microorganisms. This is an estimated 84% of all known species.  From the perspective of LinkedEarth, unfortunately this data is not currently available as linked data, although the organization does have plans to provide linked data as part of future funding applications (P. Schalk pers. comm. Nov 2015).
 +
 
 +
 
 +
== Recommendations for new datasets ==
 +
 
 +
Recommendations for collecting dendroarchaeological data have been made by Brewer and Jansma<ref>Brewer, P.W. and Jansma, E. (2016) Dendrochronological Data in Archaeology: A Guide to Good Practice, Archaeology Data Service / Digital Antiquity v1.1. [http://guides.archaeologydataservice.ac.uk/g2gp/Dendro_Toc Link]</ref>.  These are specific to the collection of samples for cultural dendrochronology, but still relevant in part to wider tree-ring research.
 +
 
 +
 
 +
 
 +
== Recommendations for legacy datasets ==
 +
The task of standardizing and upgrading legacy datasets cannot be underestimated. Recommendations for how to handle datasets will likely vary depending on the context.  For instance, if a dataset owner wants to upgrade their own legacy data they should have access to the vast majority of data/metadata.  Given the time/money they should be able to upgrade to a similar quality as new datasets.  In contrast, upgrading the existing NOAA/ITRDB holdings of others will be far more challenging as the majority of datasets are missing key information or key information is stored ambiguously.  For orphaned datasets it's very likely that a substantial amount of information is already lost with no hope of recovery.
 +
 
 +
In general, legacy datasets should be upgraded to meet a minimum set of criteria:
 +
# Type of data: raw measurements or derived chronologies
 +
# Measurement variable stored
 +
# Measurement units stored
 +
# Missing rings handled clearly and consistently
 +
# Taxon
 +
# Location
 +
# Location type
  
 +
Legacy datasets of isotopic data, in addition to the above should at a minimum include:
 +
# The type of isotope analyzed
 +
# Pre-treated or not
 +
# Method of pre-treatment
 +
# Pooled or single ring/single tree
 +
# Portion of ring analyzed (Whole ring/earlywood/latewood/middle ring)
  
 
==References==
 
==References==
 
<references \>
 
<references \>

Latest revision as of 22:03, 25 April 2017


Overview

In the Linked Earth context, a working group (WG) is a self-organized coalition of knowledgeable experts, whose activities are governed herewith. This page is dedicated to the discussion of data and metadata standards for trees (see this page for a definition of the wood archive), and aims to formulate a set of recommendations for such a standard. Note that chronological aspects should be discussed within the Chronologies WG.

Members of 'Trees Working Group'

Specific tasks

We recommend that discussions focus on the following techniques, and explore potential commonalities.

For each chronology type, we recommend:

  • structuring discussions around what scientific questions one would want to ask of the data
  • listing essential, recommended, and optional information for:
    • the measurements themselves
    • any inference made from the measurements (e.g. calibration to temperature)
    • the underlying uncertainties, and what those numbers correspond to (e.g. 1-sigma or 2-sigma?)
  • provide an ideal data table for each type of observation, so the community knows what to report and how to report it.
  • provide separate recommendations for new and legacy datasets

Scope

Tree rings are used by many different researchers, for many different reasons. The LinkedEarth community is well rooted in the paleoclimate community and while the ontology and data standard has great potential for use outside this community, we need to decide how to address the overlap in our discussions. Should be concentrate on the needs of the paleoclimate community exclusively to begin with, and intend to go back and revise at a later date? Or should we attempt to accommodate all potential users of tree-ring data from day 1?

The discussion within this group can build upon the work done by the Tree Ring Data Standard (TRiDaS) consortium [1]. While the contributors were weighted towards dendroarchaeologists, the standard was intended to be universal for all uses of tree-ring data. The requirements laid out by the TRiDaS contributors are therefore going to largely be relevant to discussions here.

Note: Do not edit the polls (even for typos) once voting has started as it will reset the vote counts to zero. If a change needs to be made, make an annotation above the old poll (i.e., above the poll tags) and place the new poll below the first one.

What should the scope of this working group be?
You are not entitled to vote.
You are not entitled to view results of this poll.
There were 6 votes since the poll was created on 16:34, 10 October 2016.
poll-id D7AF922048B5799C67A469A0BF294059

Sensors

An initial recommendation is to focus on different sensors:

Measurement variables

The traditional variable measured in dendrochronology is of course the whole ring-width. It is common now to collect a wealth of other variables from tree-rings including:

  • Whole ring width
  • Early wood width
  • Late wood width
  • Whole ring density
  • Early wood density
  • Late wood density
  • Maximum density
  • Latewood percentage
  • Vessel size
  • Blue intensity (density proxy)
  • Stable isotopes

Event information

Tree-rings provide the potential to store information about specific events. The most common of these is a forest fire (stored as a fire scar or other anatomical damage), but may include the effects of defoliating insects, volcanic eruptions etc. This sort of event data is represented as binary yes/no occurrences in particular rings and is quite distinct from the continuous measurement variables (e.g. ring-widths, isotope concentrations etc) that are typical with most proxies.


Tree rings

Detrending

Maintaining references back to raw ring-width measurements is important to maintain data transparency. When detrending has been performed a detailed record of what has been done should be stored along with the software/library used. Implementations of different detrending techniques in different software can produce different results. To ensure results can be replicated the following fields will be necessary in addition to the detrended data:

  • Software used
  • Version used
  • Detrend method name:
    • Polynomial
    • Exponential
    • Moving average / Floating
    • High pass filter
    • Cubic spline
  • In the case of some methods (e.g. polynomial, moving average) then a parameter may be required. I can only think of methods that require one parameter, but for future proofing it would make sense to be able to pass multiple parameters.

Stable isotopes

Some recommendations were made for an 'International Tree Ring Isotope Databank' ITRIDB by Csank[2]

With tree-ring isotopes workers recognized that in addition to the information provided with a ring width dataset there is a need to provide several extra pieces of information as metadata . In addition the usual information on site name, investigator, publication information, genus/species, latitude/longitude, etc. the information that was deemed important to include with any tree-ring isotope submission is listed below.

  • Whether or not a chronology was established
  • If there is no chronology (or it is floating in the case of sub fossil material) how was age established.
  • If it is tropical tree data how was the age model established and what are the uncertainties associated with that age model.
  • Were samples pooled (either single year/single tree) or are samples single ring/single tree measurements that have been averaged
  • What portion of the ring was analyzed (e.g. was it whole ring, earlywood, latewood middle ring)
  • Were were samples pre-treated by conversion to alpha cellulose, holo cellulose or in the case of 2H nitrated cellulose or was unprocessed wood analyzed
  • If pre-treated, what method of pre-treatment was used (e.g. Leavitt and Danzer, 1993; Loader et al. 1997; Brendel et al. 2000; modified Brendel (Anchukaitis et al. 2008); Rinne et 2005; Li et al. 2011; Weiloch et al. 2011; Kagawa et al. 2015)
  • What was the analysis date
  • Which isotope(s) was(were) measured (e.g. 2H, 13C, 18O, 15N, 34S)
  • Analytical precision for each isotope analyzed
  • If ẟ13C data is presented was it corrected for the Suess effect or not and if so which correction was used (e.g. McCarrol et al. 2009) (one suggestion is that only uncorrected data should be presented and users can correct their own data as is done for 14C data)
  • Which stable isotope laboratory performed the analyses.

Location information

Prior to the wide availability and affordability of GPS handsets, location information in dendrochronology tended to be low accuracy coordinates for a particular study site. This is what is stored in the ITRDB. However, for many years now the many (majority? all?) researchers have collected GPS coordinates for each individual tree. A clear distinction is needed in the ontology to record what the location information is recording.

Chronologies are by definition an amalgamation of data from multiple trees. Location information for these can therefore be represented either as multiple coordinates (one for each tree) and/or a polygon denoting the extent of the area covered by the chronology. Should such a polygon be the smallest area encompassing all tree coordinates or should it be the freehand area covering the area the researcher feels the chronology represents?

Further complications arise when handling tree-ring data collected from non-living trees. Sample from snags can be reasonably expected to be at least close to where the tree grew, but archaeological and other cultural samples can be found miles from their initial growth location. Very long range transport/trade of wood has taken place for millenia. Some studies have successfully determined the provenance of transported/traded samples through dendrochronological analyses and/or isotopic work. In these cases it would be good to also be able to store this sort of information. Regardless though, the type of location must be clearly stored with all locality data.


Taxonomic information

Taxonomic information is clearly essential for tree ring (and other biological proxies). The current most comprehensive database of authoritative taxonomic information is the Catalogue of Life. As of 2016 it contains over 1.6m species from 158 contributing databases covering plants, animals, fungi and microorganisms. This is an estimated 84% of all known species. From the perspective of LinkedEarth, unfortunately this data is not currently available as linked data, although the organization does have plans to provide linked data as part of future funding applications (P. Schalk pers. comm. Nov 2015).


Recommendations for new datasets

Recommendations for collecting dendroarchaeological data have been made by Brewer and Jansma[3]. These are specific to the collection of samples for cultural dendrochronology, but still relevant in part to wider tree-ring research.


Recommendations for legacy datasets

The task of standardizing and upgrading legacy datasets cannot be underestimated. Recommendations for how to handle datasets will likely vary depending on the context. For instance, if a dataset owner wants to upgrade their own legacy data they should have access to the vast majority of data/metadata. Given the time/money they should be able to upgrade to a similar quality as new datasets. In contrast, upgrading the existing NOAA/ITRDB holdings of others will be far more challenging as the majority of datasets are missing key information or key information is stored ambiguously. For orphaned datasets it's very likely that a substantial amount of information is already lost with no hope of recovery.

In general, legacy datasets should be upgraded to meet a minimum set of criteria:

  1. Type of data: raw measurements or derived chronologies
  2. Measurement variable stored
  3. Measurement units stored
  4. Missing rings handled clearly and consistently
  5. Taxon
  6. Location
  7. Location type

Legacy datasets of isotopic data, in addition to the above should at a minimum include:

  1. The type of isotope analyzed
  2. Pre-treated or not
  3. Method of pre-treatment
  4. Pooled or single ring/single tree
  5. Portion of ring analyzed (Whole ring/earlywood/latewood/middle ring)

References

  1. Jansma, E., Brewer, P. W. & Zandhuis, I. (2010) TRiDaS 1.1: The tree ring data standard. Dendrochronologia 28, 99-130. DOI
  2. Csank, A.Z. (2009) An International Tree-Ring Isotope Data Bank– A Proposed Repository for Tree-Ring Isotopic Data. Tree-Ring Research 65(2):163-164. DOI
  3. Brewer, P.W. and Jansma, E. (2016) Dendrochronological Data in Archaeology: A Guide to Good Practice, Archaeology Data Service / Digital Antiquity v1.1. Link

This category currently contains no pages or media.