P. Brewer & E. Jansma, 2015: Dendrochronological Data in Archaeology: A Guide to Good Practice

Share Embed


Descrição do Produto

Dendrochronological Data in Archaeology: A Guide to Good Practice1 Peter Brewera & Esther Jansmab

a. Laboratory of Tree-Ring Research , University of Arizona , USA b. Cultural Heritage Agency and Utrecht University , The Netherlands2

VERSION 1.0.0 - 1 JUNE 2015

Some wooden objects dated through dendrochronology3

1

This text has been derived from http://guides.archaeologydataservice.ac.uk/g2gp/Dendro_Toc (website consulted on October 9th 2015). 2 Other organizations involved in the development of this Guide are: Archaeology Data Service (ADS; UK), Data Archiving and Networked Services (DANS; NL) and EU 7th-framework programme ARIADNE. 3 Image derived from: Jansma, E. & R.J. van Lanen (2012) ‘Een digitale bibliotheek van dateringen: de internationale doorwerking van een Nederlands initiatief’. Vitruvius 20, 36-41.

1

Content Section 1 - Aims and Objectives.............................................................................................. 3 1.1 Background to the Guide's Focus in an Archaeological Context..................................... 3 1.1.1 What is Dendrochronology?........................................................................................ 3 1.1.2 Applications of Dendrochronology in Archaeology ................................................... 3 a) Obtaining absolute calendar dates for archaeological sites/structure ................ 3 b) Establishing timber provenance and long-distance relations............................. 4 c) Assessing aspects of construction technology ................................................... 5 d) Reconstructing environmental history ............................................................... 5 1.1.3 Current Issues or Concerns ........................................................................................ 5 a) The sustainability of dendrochronological data ................................................. 5 b) The reliability of dendrochronological data and results .................................... 5 c) Digital formats of dendrochronological data ..................................................... 6 d) The need for additional (meta)data .................................................................... 6 1.2. Scope of the Guide .......................................................................................................... 6 1.3 Data and Metadata............................................................................................................ 7 Section 2 - Creating dendrochronological data ..................................................................... 7 2.1 Project Planning and Requirements ................................................................................. 7 2.2 Sources of Data ................................................................................................................ 8 2.3 File Types (whilst creating, working with, processing data) ........................................... 8 2.4 File Naming Convention .................................................................................................. 9 2.5 Documenting Data Creation and Processing.................................................................. 11 Section 3 - Archiving dendrochronological data ................................................................. 11 3.1 Deciding What to Archive.............................................................................................. 11 3.2 Deciding How to Archive .............................................................................................. 11 3.3 Archiving File Types...................................................................................................... 12 3.4 Converting Data Formats ............................................................................................... 14 3.5 Archiving Strategies ....................................................................................................... 15 3.6 Metadata and Documentation......................................................................................... 15 Section 4 - Copyright.............................................................................................................. 17 References ............................................................................................................................... 18 Bibliography......................................................................................................................... 18 Resources ............................................................................................................................. 19 Dendro Data Repositories ............................................................................................. 19 Software and Data Standards ........................................................................................ 19 Glossary................................................................................................................................... 20

2

Section 1 - Aims and Objectives 1.1 Background to the Guide's Focus in an Archaeological Context 1.1.1 What is Dendrochronology? Dendrochronology is applied in cultural-heritage research including archaeology to determine the exact calendar age of ancient wood. Such age determinations contribute significantly to assessments of the meaning of archaeological and architectural structures in terms of their chronological and cultural context. This method uses the fact that in climate zones with distinct growing seasons (i.e. outside the tropics and subtropics) trees respond to seasonal temperature differences with a growth stop in autumn, after which growth starts again in spring. This seasonal rhythm is laid down in annual growth rings. The width of each ring reflects the environmental conditions during the growing season, such as temperature, precipitation and soil conditions, as well as local impacts such as flooding, fire and forest clearing or thinning. The alteration of wide and narrow growth rings in ancient wood provides a key to the exact period during which this wood was formed. As an absolute dating method dendrochronology is restricted to the last 12,000 years (Holocene), although the availability of reference chronologies means that in some regions dates are only possible for more recent time periods. Obtaining absolute dates through dendrochronology results in tree-ring measurement series that, when combined, are useful for the reconstruction of former trade/exchange networks, the former landscape and its uses, wood technology and other topics. Therefore it is essential that measurement series are deposited in trusted repositories and made available for follow-up research. 1.1.2 Applications of Dendrochronology in Archaeology For a great deal of human history, wood has been an important construction material. Remnants of ancient wood are preserved to this day in archaeological sites on land and under water, as well as in buildings and mobile heritage. Dendrochronology is applied to this wood with the following purposes: a) Obtaining absolute calendar dates for archaeological sites/structure When a piece of wood has been dated using dendrochronology, we know the exact calendar year during which each of its rings was formed. Absolute dates are obtained by comparing measurement series of undated wood patterns to absolutely dated reference chronologies of average annual tree growth. There are several manners in which the growth patterns of archaeological wood are measured, ranging from destructive (using a cross section or thin core taken out of a timber) to nondestructive (measurements taken from the outside of an artefact). When a cross section is available, the surface is prepared in such a manner that the annual rings become visible, after which ring widths are measured directly from the sample or (on screen) from a photo using a 0.01 or 0.001mm resolution. During this process in most cases dendrochronological measuring tables are used which are linked to a computer. When measurements are taken from the original wood surface, either photos or a hand lens are used. In the latter case measurements are registered on paper and are entered into the computer afterwards.

3

When dating archaeological wood a wide variety of reference chronologies are needed, because: (i) tree growth reflects environmental conditions that vary regionally and even - in highly variable landscapes - across small distances; (ii) different tree species respond differently to environmental conditions; (iii) the timber may have been brought to the construction site from other regions. Comparisons to reference chronologies are performed using a variety of commercial and open source software. Dendrochronological dates are based on statistical variables that express the agreement between time series, on replication of results (significant matches against a variety of chronologies), and on visual verification of statistical agreements (expert judgement). The calendar year during which the outer ring of an archaeological timber was formed does not necessarily provide an absolute date for the archaeological structure in which this timber was used. Reasons for differences between a dendrochronological and construction date are: (i) the wood is incomplete, i.e. lacks the outer rings; (ii) the wood was stock piled before use (i.e. used some years after it was harvested); (iii) the wood was re-used (i.e. had a previous function); (iv) the wood represents a later repair. The first issue (i) is dealt with in dendrochronology by estimating a minimum number of missing outer rings and using this number to derive a terminus post quem date for the felling of the tree. Not all wood can be dated by dendrochronology. Some species of tree are unsuitable for dendrochronology because they do not lay down reliable annual growth increments. Even species that do typically have annual growth increments may not be successfully dated if localised factors that affected the width of the rings (e.g. injuries; growth releases caused by the removal of competing trees) mask the regional signal used to discern matches between tree-ring patterns. Samples must also include sufficient rings to enable a secure match to be found with reference chronologies therefore short lived trees and branches are not suitable. Samples containing less than 60 rings as a rule are unsuitable for dendrochronological dating. Note, however, that some trees grow very slowly and therefore the size of the sample is not indicative of its usefulness for dendrochronology. Samples with small diameters can contain sufficient rings to date them successfully whereas very large samples sometimes contain very few rings. Also note that some tree species require lower ring numbers for dendrochronological dating than other species. In addition the number of rings that is needed depends on the severity of the environmental conditions that ruled the trees growth. The assessment of dendrochronological quality in terms of tree species and ring numbers is best left to a dendrochronologist experienced in the research of wood from the study area. b) Establishing timber provenance and long-distance relations The provenance of archaeological wood is determined based on the geographical provenance of the reference chronologies that provide the best dendrochronological matches. As the result of recent ICT developments in dendrochronology a data-network approach has become more common when assessing wood provenance. This application is useful when studying once mobile artefacts such as ships and barrels, to assess the scale and organisation level of large infrastructural works in open and deforested landscapes, and to reconstruct former trade and exchange networks. Oak is most commonly used in European provenance studies since oak timber was a valuable commodity in the past due to its structural and chemical properties and due to the existence of large international data collections of this species.

4

c) Assessing aspects of construction technology Using dendrochronology it is possible to establish that separate timbers were derived from the same tree. Identifying trees in archaeological wood assemblages allows the reconstruction of aspects of, among others, former cooperage and ship-building technologies. d) Reconstructing environmental history Dendrochronology is possible because the annual variation in environmental conditions causes differences in the width of consecutive tree-ring widths. The type of environmental changes that trees respond to depends on a number of factors including the species and the position of the tree within its ecological range. For example, a species adapted to dry environments may be limited in some regions by excessive precipitation, while in another region the same species may be limited by winter temperatures. Statistical analysis of the ring widths can extract the environmental signals that limited growth, thus enabling the reconstruction of past temperature and precipitation and, in low-lying delta regions, of flooding due to climate and/or man-induced deforestation. 1.1.3 Current Issues or Concerns a) The sustainability of dendrochronological data Locally-managed dendrochronological collections are lost through the disposal of analogous collections, retirement of researchers and the closing down of research departments and commercial laboratories. This is why an international consortium of organisations has created the European heritage-based digital repository DCCD4 (Jansma et al. 2012a; Jansma 2013). Dendrochronological collections are typically at much greater risk of disposal than other archaeological collections due to the inherent value of wood samples being much less obvious. Dendrochronological samples and their associated data and metadata should be sustainably curated to enable the re-evaluation of existing datasets and the generation of new datasets as techniques become available. b) The reliability of dendrochronological data and results Dendrochronology among others is offered commercially by (international) suppliers who may operate outside national systems of cultural-heritage quality regulations. This poses a risk to scientific accuracy, data sustainability and follow-up research. A common example of scientific inaccuracy is the presentation of calendar dates for tree-ring series that are too short to be dateable using dendrochronology. Even a long sequence of ring-widths does not guarantee a cross match will be possible. Less reputable laboratories may report the 'best' match for a sample as a calendar date, regardless of the significance of the fit. It is therefore essential that users of dendrochronological services insist upon the delivery of a comprehensive report along with the calendar date(s) for the submitted samples. Most importantly the report should include: the raw ring-width data; the reference chronologies (or a full description of the chronologies if they are not freely available) used as the basis for the dating; t-score and/or %PV (Gleichlaufigkeit) plus level of significance statistics for the cross-date; and graphs showing the visual similarity between tree-ring patterns and reference chronologies. Results produced by laboratories that are unwilling or unable to provide such 4

Digital Collaboratory for Cultural Dendrochronology (DCCD) - http://dendro.dans.knaw.nl

5

supporting information should be treated with caution. As with any scientific analysis, dendrochronological determinations should be subject to review and critique by independent third parties where necessary. c) Digital formats of dendrochronological data Dendrochronology has a long history of utilizing computers to assist with research. This has resulted in a large number of digital data formats each with their own strengths and weaknesses. The vast majority of these formats concentrate on the storage of tree-ring measurement series with little concern for metadata (see section 1.3 for a definition of data versus metadata in dendrochronology). Even the storage of ring-width data within many of these formats presents issues. For example many formats do not explicitly include measurement units and others do not allow for the recording of missing rings (i.e. years in which the conditions were so bad that the tree didn’t add a ring at all). The large number of data formats also present technical issues with the sharing of data, since many formats are only supported by the original software that they were designed to be used with. To address these and many other issues, an international consortium of dendrochronologists, computer scientists and users of dendrochronological data have created the Tree-Ring Data Standard (TRiDaS)5 [2] (Jansma et al. 2010). d) The need for additional (meta)data In many scientific fields the research questions determine which (combination of) variables are studied, which often leads to the necessity to collect new data according to a specific research plan. Dendrochronology differs in this respect because it always uses tree-ring width as a main variable. This is why for many applications of dendrochronology existing tree-ring measurement series can be used to answer (new) research questions (see section (case studies). The suitability of dendrochronological measurement series for new research depends on the available information about the series, such as the site or object that contained the wood, the dendrochronological dates and the tree species. This is why the inclusion of rich and associated (meta)data is essential when archiving dendrochronological measurement series The DCCD repository at present is the only tree-ring archive worldwide that allows for the standardised inclusion of such (meta)data.

1.2. Scope of the Guide This document serves as a good-practice guide for the collection and archival of dendrochronological data in the context of archaeological and historical research. The guide is aimed at both those creating dendrochronological datasets, and those that commission dendrochronological analyses. This guide does not cover the methods involved in dendrochronological analyses, but focuses on how to describe and archive the data and metadata involved in these analyses. This guide is concerned with best practice for the curation of digital information but does not cover the equally important aspects of the curation of physical samples. However, physical samples are the primary source of information in dendrochronological analyses and should always be managed alongside the digital data wherever possible. This ensures that samples can be re-evaluated where necessary and also reexamined as new analytical techniques are developed.

5

Tree Ring Data Standard (TRiDaS) Website - http://www.tridas.org/

6

It is also noted that as this guide is concerned with the management of digital dendrochronological data, it does not cover the issues associated with archiving data produced solely via the 'skeleton plotting' and associated methods of dendrochronology. Skeleton plotting is a technique developed by Douglass in the early 20th century and still commonly used, especially in the USA, whereby trees are crossdated using hand-drawn plots on graph paper. Skeleton plotting can be used to provide a dendrochronological date without the generation of any digital data, although it is common to measure samples even after skeleton plotting to enable the calculation of statistics and to facilitate further analyses.

1.3 Data and Metadata In many archaeological archives there is a clear difference between data and metadata (data about data). Making this distinction in dendrochronology, however, can be quite difficult. It is important to make a clear distinction firstly for the sake of clarity within this document (this is particularly necessary in section 3.3 for example) but also to assist in wider discussions surrounding data use and intellectual property. All the information and knowledge generated by the dendrochronologist during the course of their analysis is regarded as data. The most important data are the measurement values for each ring (typically whole ring widths but it can also be a variety of other metrics). In addition the following data are essential since they are required for the interpretation of measurement series: sapwood counts; species; presence of bark, and calendar dates. Although these fields may be considered as describing the sample being analysed, they are derived during the analysis process itself and as such are regarded as data. In contrast, metadata is typically provided by the archaeologist commissioning the analysis, or is collected by the dendrochronologist in the field at the time of sampling. Examples include: location; trench number; find number; artefact type, etc. The exceptions to this rule are associated documents such as publications, reports, photographs, etc., which are also considered metadata. Note though that these will typically also contain data.

Section 2 - Creating dendrochronological data 2.1 Project Planning and Requirements Project planning requirements will vary dramatically depending on many factors including: the type of object(s); state of preservation; species; size of samples; whether destructive sampling is allowed; financial resources. What and how to sample should be carefully considered following standard dendrochronological procedures (table 1) while making informed decisions about what metadata to collect. Table 1 Examples of national guidelines for dendrochronological analyses. Country Author Year Title NL

Jansma

2002 Veldhandleiding dendrochronologisch onderzoek

NL

Jansma

2006

Nederlandse onderzoeksagenda voor archeologie: Dendrochronologie

7

UK

Historic England

1998

Dendrochronology: Guidelines on producing and interpreting dendrochronological dates

IT

UNI

2004

Cultural heritage - Wooden artefacts - Wood dendrochronological dating guidelines

Metadata should follow the Tree Ring Data Standard (TRiDaS) data model which provides the ability to record a wealth of information. Not all TRiDaS fields will be applicable at any one time, but consider recording as much metadata as possible without adding an undue burden to the project. If you have very simple research goals you may need very little metadata to achieve your aims, but with the addition of just a few metadata fields, it is likely your dataset will be of much greater use to you and fellow researchers in the future.

Figure 1: Diagram show the general structure of the Tree Ring Data Standard (TRiDaS) data model. For further description of the component parts of the model see the glossary.

2.2 Sources of Data Dendrochronological data is typically generated either by the measurement of rings using a standard measurement platform (e.g. Velmex, Lintab, Catras, etc.) or via the manual measuring of rings with a hand lens. Using either method, the resulting data typically ends up in specialist dendrochronological software for analysis.

2.3 File Types (whilst creating, working with, processing data) There are currently approximately 24 different dendrochronological data formats in use within the dendrochronological community. While most are adequate for creating and analysing

8

dendrochronological ring-width series, many of these formats are read by just one or two programs making them unsuitable for archival or data sharing. Probably the most commonly used format worldwide is the Tucson decadal format which has been used since the late 1970s. It’s a relatively simple ASCII-based format based upon the structure of the punch cards it replaced. Whilst very common in laboratories around the world, the format has a number of limitations, not least the fact it has almost no capacity to store associated metadata. Limitations with regards the storage of metadata within the format, and the lack of a formal standardisation procedure have meant that a variety of ad hoc methods have been introduced by users and software developers over the decades. These result in digital ambiguities and incompatibilities. Other popular formats, especially in Europe, are Heidelberg and CATRAS. The CATRAS format was designed for the program of the same name (Aniol, 1983, 1987), while the Heidelberg format was developed for use with TSAP (Rinn 2005). Both formats have some support in some other dendrochronological applications, but this support varies, especially with regards their less popular features. The CATRAS format is an undocumented proprietary binary format and as such we strongly recommend against using it for archiving or data sharing. In comparison, Heidelberg files are plain-text and the format is relatively well documented. While still not ideal for archival purposes, Heidelberg files are suitable for shortterm storage and sharing of data. The Tree Ring Data Standard (TRiDaS) was introduced in 2010 as a universal format for describing any sort of tree-ring data or metadata. We strongly recommend using the TRiDaS format for the long-term storage of dendrochronological data and metadata. One major obstacle to the universal adoption of TRiDaS is the availability of applications and utilities for processing and working with TRiDaS format data. At the time of writing this guide, the only application suitable for day-to-day collection, analysis and management of native TRiDaS data is Tellervo (Brewer, 2014; Brewer et al 2010). A number of applications routinely used within the dendrochronological community have their roots back in the 1980s (e.g. COFECHA and DPL). While these products are venerable, they still perform well and therefore until improvements or replacements can be made available, dendrochronological workflows must adapt to their limitations. With this in mind the universal dendrochronological data conversion tool - TRiCYCLE - was developed soon after the release of the tree ring data standard (Brewer et al. 2011). See section 3.3 for further information.

2.4 File Naming Convention Folder structure and file naming conventions should be decided upon and well documented before work begins. Documentation should be kept in the base of the folder structure for reference. For ease of access we strongly recommend a folder structure that follows the TRiDaS data model. For instance a top level folder for each project, followed by one subfolder for each 'object' (site), and nested sub-folders to handle sub-objects if necessary. Within these folders are folders for elements, samples etc. The folder structure may be simplified and stopped at sites or trees, but in such cases it is essential to explicitly follow the file naming scheme outlined below.

9

Figure 2: Recommendation for structure of folders when saving dendrochronological data files. Note the structure follows the TRiDaS data model with the top level folder below the root of the repository corresponding to a TRiDaS project.

File naming conventions should also follow the TRiDaS data model. This, however, can be problematic if you need to maintain compatibility with legacy (typically DOS-based) programs that have restrictions on file name length. DOS programs can only handle file names with 8 characters plus a 3 character file extension. If 8 character names are required then we recommend: • • • •

3 alpha character code for object e.g. ABC 3 digit code for element e.g. 001 1 character code for sample e.g. A 1 digit sequential number to distinguish between files

Note that this unfortunately necessarily excludes the concept of a TRiDaS radius. This would result in a Tucson file being named: ABC001A1.rwl. This naming convention can also serve as a guide for the concept of a 'keycode'. A number of legacy file formats use the concept of an 8 character 'keycode' for identifying a series. When using these short file names it is more important to maintain the deep nested folder structure described above. If backwards compatibility with DOS programs is not required, then the file naming convention should be longer and more explicit. We recommend delimiting the TRiDaS entities within the file name using the hyphen and to include all levels of the TRiDaS data model. For example: ABC-1-B-A-1.rwl would refer to the first measurement series, from radius A, of sample B, of element 1 from object ABC.

10

We do not support the use of the file extension characters for anything other than identifying the format of the file. This is a widely understood use for these characters and altering them to compensate for file length limitations can lead to confusion. File extensions should follow established conventions e.g. .fh for Heidelberg, .rwl for Tucson etc. If the format does not have a widely agreed upon file extension, then .txt should be used instead. If deviations are necessary from the naming and structuring proposals outlined above, then it is essential that detailed documentation be made and stored with the data files. These conventions must be adhered to, to ensure that those that follow can understand exactly what each data file refers to.

2.5 Documenting Data Creation and Processing TRiDaS contains the capability to describe all aspects of the creation and processing of digital dendrochronological information. The software and methods used to index and cross-date dendrochronological data sets can and should be included wherever possible. The support for such extensive use of TRiDaS in software is currently very limited.

Section 3 - Archiving dendrochronological data 3.1 Deciding What to Archive In comparison to many of the other data types covered by this good practice guide series, dendrochronological data files are very small and therefore the burden of retention is significantly lower. It should be noted, however, that researchers are increasingly storing high-resolution scans and photos of wood surfaces which can substantially increase the space required for archiving. Even with this consideration, it is good practice to archive all verified dendrochronological data and metadata including photos and research reports. For guidance regarding best practice for the creation and archival of image files, see the Raster Images Guide. What can be archived varies due to the policies of the repository being archived to. For example the International Tree Ring Data Bank6 (ITRDB) has strict data-quality control mechanisms resulting in the storage of verified and accurate data sets. The DCCD repository does not implement quality control which implies that the quality of the datasets stored in this repository is unverified and the responsibility of the author(s) of the data.

3.2 Deciding How to Archive Dendrochronological data should be made publicly available wherever possible in order to enable follow-up research and to assist cultural-heritage management. Many data sets are needed to build the chronologies necessary to date archaeological and historical structures, which means that the entire dendrochronological community benefits from the open sharing of data. In addition the ancient world was very interconnected, and often researchers will need the assistance of dendrochronologists working in other regions to accurately date wood 6

International Tree Ring Data Bank (ITRDB) - https://www.ncdc.noaa.gov/data-access/paleoclimatologydata/datasets/tree-ring

11

samples. Hence it is only by collaboration and effective data sharing that we will get the most from dendrochronology. There are currently two internationally recognised repositories for tree ring data: the International Tree Ring Data Bank4 (ITRDB); and the Digital Collaboratory for Cultural Dendrochronology7 (DCCD). The ITRDB is hosted by the National Oceanic and Atmospheric Administration (NOAA) in the USA and is primarily aimed at dendrochronological data for climate research although it does also contain archaeologically-sourced dendrochronological data. The repository was set up following an international workshop in 1974 and is structured around the widely used Tucson decadal data format. The ITRDB integrates with a number of other palaeo-climate proxy databases therefore the metadata it stores is rather minimal and generic. The ITRDB does support the submission of additional data files so TRiDaS data should be submitted where possible. The DCCD is a newer repository launched in 2012 hosted and maintained by Data Archiving and Networked Services8 (DANS) in the Netherlands. The repository is based on TRiDaS and is primarily aimed at dendrochronological data associated with cultural research including archaeology, architecture, art, etc. While DCCD users are encouraged to make their data publicly available, the repository has the facilities to enable the data provider to control access permissions. This repository is therefore a good choice for users who do not wish - or are unable to make - their data publicly available yet and who meanwhile want to ensure their data is securely managed and maintained in a recognised repository. This procedure ensures that data that is orphaned by the death of researchers or the closure of departments or companies is not lost. In some circumstance, perhaps for regulatory reasons, data needs to be stored in other designated generic repositories. It should be noted that in such cases the usefulness of the data for follow-up research will likely be limited. Generic repositories are more suited to allow the re-examination or verification of specific dendrochronological analyses, rather than the reanalysis of data in new ways (e.g. regional, national or international syntheses). Even when required to archive data into other designated archives, it is also recommended to submit data to one of the specialist dendrochronological archives described above. If given the option of choosing a generic repository, care should be taken to ensure it is backed by the international EU 'Data Seal of Approval' or a similar authority, which indicates the archive is structured and managed according to current best practices. Any dendrochronological data archived in generic repositories should be stored in TRiDaS format files.

3.3 Archiving File Types As mentioned in section 2.3, there are a great many data formats in use within the dendrochronological community. However, with the exception of the Tree Ring Data Standard, none provide the mechanism for recording rich standardised metadata. Wherever possible we strongly recommend using TRiDaS format files to store both your data and metadata.

7 8

Digital Collaboratory for Cultural Dendrochronology (DCCD) - http://dendro.dans.knaw.nl/ Data Archiving and Networked Services (DANS) - http://www.dans.knaw.nl

12

If the tools being used for the dendrochronological analyses do not support TRiDaS, then one option is to maintain your data within the legacy format supported by the software, and the metadata in a TRiDaS-enabled database like TRiDaBASE[4] (Jansma et al. 2012b). The DCCD repository supports the submission of TRiDaS plus a variety of legacy formats, whereas the ITRDB supports the submission of Tucson and TRiDaS files. The managing of the legacy data files in a hybrid TRiDaS/legacy system is definitely the weakest link in this approach. With no software to provide error checking, it is all too easy to mislabel, rename, move and delete files - especially in a multi-user environment. The utmost care must be taken at all times to be as strict as possible. Write access to the files should be limited to only essential personnel. Folders containing site data files should be made readonly as soon as work on a site has been completed. If one decides to follow this method it is important to ensure that the legacy data format that is used stores the ring-width data in an unambiguous manner. Where ambiguities exist, these should be clarified in the TRiDaS metadata. A list of formats, along with some of the potential issues they may have, is available in table 2. Further details are available in Brewer et al. (2011). One issue shared by many of the legacy data formats is the ambiguity of the calendar used for dates. To reduce the programming complexity, many formats use what is known as the astronomical rather than the Gregorian calendar. The astronomical calendar includes a year zero and denotes BC dates as negative integers. This means that prior to the AD/BC transition the year numbers are offset by one (e.g. 1AD = 1; 1BC = 0; 2BC=-1; etc.). This can cause complications for researchers both outside and within the dendrochronological community. It is not uncommon to find legacy data files with BC data where the offset has been removed. Table 2 - Legacy files formats: assessment of their suitability for storing dendrochronological data in combination with a companion TRiDaS file containing metadata. Suitable for Format TRiDaS coPotential issues archival Belfast Apple

With caveats Format cannot store missing ring information

Belfast Archive With caveats Format cannot store missing ring information Besancon

The format should strictly only contain data values are in 1/100th mm, however, some users store micron resolution With caveats data instead. There is no standard way to discern the difference

CATRAS

No

Closed source proprietary binary format. There is some support for reading/writing CATRAS files in the TRiCYCLE library but the closed nature of the format means this is not comprehensive

Cracow Binary format

No

Simple binary format storing just ring-widths and sapwood markers. Although this is a very simple format, the fact it is binary means accessing the data is not trivial

Corina

No

File format with very limited software support

DendroDB

No

An export format used by the discontinued DendroDB

13

database. No known software implementations generate this format. No

Dendro format for recording fire history event data. Does not support ring-width data

No

Various spreadsheet-style formats (including CSV, MS Excel, ODF etc) are used by some to store and transfer ringwidth data. The flexible nature of spreadsheets means that data can be stored in a wide variety of ways makes it unsuitable for long term storage of data.

Heidelberg

Yes

Native format for the widely used TSAP software. Has extensive support for metadata fields, but the contents of these fields are not standardised

KINSYS-KS

Format specific to the software of the same name produced by the Finnish Forest Research Institute although not widely With caveats used elsewhere. Stores ring-width data effectively but the lack of support in software means it is not ideal for long term storage

Nottingham

No

Oxford

With caveats Does not store data: prior to 1AD; ring-widths >=1mm

PAST4

No

Sheffield

With caveats Format cannot store missing ring information

FHX2 Generic spreadsheet formats

SYLPHE

A format with no known documentation or extant reference implementation Files can be in either astronomical or Gregorian calendar which can introduce ambiguities. (see Bescancon) Very simplistic format which should only be used for raw ring-width data

Topham

With caveats

TRIMS

With caveats Format cannot store missing ring information

Tucson

There are many undocumented alterations to this format. With caveats Check files with TRiCYCLE or COFECHA to ensure readability before archiving

Tucson compact No

Rather complicated format that is not widely supported

VFormat

No

Highly encoded metadata makes this format not very accessible

WinDENDRO

No

Proprietary delimited text format

3.4 Converting Data Formats All the formats listed in table 2 are supported in the universal dendrochronological data conversion tool TRiCYCLE9 (Brewer et al. 2011). TRiCYCLE is an open source, crossplatform tool available as a desktop application and also as a library for use within other programs. 9

TRiCYCLE - http://www.tridas.org/tricycle/

14

TRiCYCLE makes use of TRiDaS as an intermediary in the conversion of data and metadata to/from the 24 data formats identified within the dendrochronological community. TRiCYCLE includes a reader for each supported format that translates the data and metadata into TRiDaS, as well as a writer for each format to translate from TRiDaS to the legacy formats. This architecture means that data and metadata can be converted back and forth between any combination of formats. It should, however, be noted that depending on the formats used the process can be 'lossy' in that a round trip conversion may not result in an identical file. The ambiguous nature of many of the formats (e.g. lack of measurement units) means that TRiCYCLE is forced to make assumptions, which although clearly presented to the user, may result in some degradation of information. Users must therefore take time to understand the limitations of each format.

3.5 Archiving Strategies The TRiDaS community is committed to maintaining this dendrochronological standard, including adding support for new features as deemed necessary. The standard will, however, remain both backwards and forwards compatible ensuring easy access to existing TRiDaS data for years to come. It should therefore not be necessary to periodically refresh or migrate data to newer versions of the standard.

3.6 Metadata and Documentation Rich, standardised metadata is essential to ensure archived data can be effectively reused. The Tree-Ring Data Standard provides a robust data model around which this metadata can be stored. A full description of the TRiDaS data model along with all the fields that it supports, can be found in the original publication by Jansma et al (2010). While the more metadata the better, there may not be the time and resources to comprehensively store all metadata. This is often especially true when attempting to retrospectively archive existing collections where the metadata is not easily accessible or perhaps completely missing. To be as accommodating as possible few of the many data fields defined in TRiDaS are marked as mandatory. However, there are a number of optional fields that are considered extremely desirable when archiving dendroarchaeological data. A list of all mandatory and strongly recommend fields are provided in table 3. Table 3 - Metadata fields that are mandatory or strongly recommended when archiving dendrochronological data. Field Mandatory Definition project.title

Yes

Name of the project

project.identifier

Yes

Laboratory project identification such as a report number

project.type

Yes

Examples include: dating; provenance; wood technology, vegetation reconstruction; climate study

project.laboratory

Yes

Name of the dendrochronological research laboratory

15

project.category

Yes

One of: former vegetation; archaeology; building history; ship's archaeology; art/furniture; actual vegetation

project.investigator

Yes

Name of the principal investigator

project.period

Yes

When the dendrochronological project took place

object.title

Yes

Name of the object being analysed e.g. the name of a ship, archaeological site, building or painting.

object.type

Yes

Functional description: building (church, house etc.) water well, painting, musical instrument (and type), ship (and type), forest

object.latitude

No

Latitude of the object in WGS84 decimal degrees

object.longitude

No

Longitude of the object in WGS84 decimal degrees

element.title

Yes

Name of the element

element.taxon

Yes

The most detailed taxonomic name known: species, genus, family etc, preferably from the Catalogue of Life

element.latitude

No

Latitude of the element in WGS84 decimal degrees

element.longitude

No

Longitude of the element in WGS84 decimal degrees

sample.title

Yes

Name of the sample

sample.type

Yes

Method that was used to take a sample from the element e.g. core, section.

sample.samplingdate

No

Date the sample was taken

radius.title

Yes

Name of the radius

radius.pith

Yes

Whether the pith is present or absent

radius.sapwood

Yes

One of: n/a; absent; complete or incomplete

radius.bark

Yes

Whether the bark is present or absent

measurementSeries. title

Yes

Name of the measurement series

measurementSeries.measuringmethod Yes

What method was used to measure the sample

measurementSeries. variable

Yes

Measured variable e.g. ring-width, earlywood, latewood etc

derivedSeries.type

Yes

If the series has been derived from other series, what type of series is it e.g. chronology, object curve etc

16

Although the TRiDaS data model (i.e. the structure and data fields) is typically implemented as XML files that comply to the TRiDaS XML schema, it can also be followed in other less formal ways. For instance the Heidelberg data format enables users to store user-defined metadata fields so it is possible to use the TRiDaS fields definitions as additional user-defined fields within a Heidelberg file. If you are using the TRiDaS data model in such a manner, take time to get to know the structure of the model along with its deliberately introduced formatting restrictions. For example, TRiDaS extends the Geography Markup Language (GML) for storing location information. If you are storing coordinates, take care to format them in a logical and consistent manner and to record associated information such as the datum. TRiDaS also defines a number of dictionaries for common terms. These dictionaries should be followed wherever possible.

Section 4 - Copyright Beyond the standard issues raised in the general Guide section on Copyright and Intellectual Property Rights (IPR), there are no additional issues to raise for dendrochronological data. It should be noted, however, that dendrochronological analyses are quite unusual in their reliance on existing reference datasets. Without the open sharing of datasets, each researcher would need to undertake the painstaking task of building highly-replicated chronologies from the present day back to the time period of interest for each new region they work in. For archaeology-based wood research to continue to flourish, dendrochronologists need to continue to contribute datasets to the publicly accessible repositories. Care should be taken to investigate the copyright and IPR agreements for the repository that you are using. These vary considerably and the details of specific archives is beyond the scope of this guide.

17

References Bibliography Aniol, R.W. (1987) 'A new device for computer measurement of tree-ring widths'. Dendrochronologia 4, 135-141. Aniol, R.W. (1983) 'Tree-ring analysis using CATRAS'. Dendrochronologia 1, 45–53. Brewer, P.W. (2014) 'Data management in dendroarchaeology using Tellervo'. Radiocarbon 56, and Tree-Ring Research 70, S79-83. https://journals.uair.arizona.edu/index.php/radiocarbon/article/view/18320 Brewer, P.W., Murphy, D. & Jansma, E., (2011) TRiCYCLE: a universal conversion tool for digital tree-ring data. Tree-Ring Research 67, 135-144. http://www.treeringsociety.org/TRBTRR/TRBTRR.htm Brewer, P.W., Sturgeon, K., Madar, L. & Manning, S.W., (2010) A new approach to dendrochronological data management. Dendrochronologia, 28: 131-134 http://dx.doi.org/10.1016/j.dendro.2009.03.003 Historic England (1998) Dendrochronology: Guidelines on producing and interpreting dendrochronological dates. https://historicengland.org.uk/imagesbooks/publications/dendrochronology-guidelines/ Jansma, E. (2002) Veldhandleiding dendrochronologisch onderzoek. In A. Carmiggelt & P.J.W.M. Schulten (eds.), Veldhandleiding archeologie; archeologie leiddraad 1, CvAK, 111114. http://www.academia.edu/2576134/ Jansma, E. (2006) 'Dendrochronologie'. In Nationale Onderzoeksagenda voor de Archeologie (NOAA), version 1.0, accepted January 2006. http://archeologieinnederland.nl/sites/default/files/3%20DEF%20Jansma%20Dendrochronolo gie.pdf Jansma, E. (2013) 'Towards sustainability in dendroarchaeology: the preservation, linkage and reuse of tree-ring data from the cultural and natural heritage in Europe'. In Bleicher et al. (eds.) DENDRO -Chronologie, -Typologie, -Ökologie. Freiburg, 169-176. Jansma, E., Haneca, K., Kosian, M. (2014) 'A dendrochronological reassessment of three Roman boats from Utrecht (the Netherlands)'. Journal of Archaeological Science 50, 484-496. 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 Jansma, E. & VAN LANEN, R.J. (in press) 'The dendrochronology of Dorestad: placing early-medieval structural timbers in a wider geographical context'. In Willemson, A., & Kik, H. (red.) Golden Middle Ages in Europe. New Research into Early-Medieval Communities and Identities. Brepols Publishers.

18

Jansma, E., Van Lanen, R.J., Brewer, P.W. & Kramer, R. (2012a) 'The DCCD: a digital data infrastructure for tree-ring research'. Dendrochronologia 30, 249-251. http://www.sciencedirect.com/science/article/pii/S1125786512000367 Jansma, E., Van Lanen, R.J., Sturgeon, K., Mohlke, S., & Brewer, P.W., (2012b) 'TRiDaBASE: A stand-alone database for storage, analysis and exchange of dendrochronological metadata'. Dendrochronologia 30, 209-211. http://dx.doi.org/10.1016/j.dendro.2011.09.002 Rinn, F. (2005) TSAP-WIN. Time-Series Analysis and Presentation for Dendrochronology and Related Applications. Version 0.53 for Microsoft Windows. User reference. UNI (2004) Cultural heritage - Wooden artefacts - Wood dendrochronological dating guidelines. http://store.uni.com/magento-1.4.0.1/index.php/uni-11141-2004.html

Resources Dendro Data Repositories • •

Digital Collaboratory for Cultural Dendrochronology (DCCD). http://dendro.dans.knaw.nl International Tree-Ring Data Bank (ITRDB). http://www.ncdc.noaa.gov/dataaccess/paleoclimatology-data/datasets/tree-ring ]

Software and Data Standards • • • •

Tree Ring Data Standard (TRiDaS). http://www.tridas.org Tellervo. http://www.tellervo.org TRiCYCLE. http://www.tridas.org/tricycle TRiDaBASE. http://www.tridas.org/tridabase

19

Glossary TRiDaS Project A project is defined by a laboratory and encompasses dendrochronological research of a particular object or group of objects. Examples include: the dating of a building; the research of forest dynamics in a stand of living trees; the dating of all Rembrandt paintings in a museum. What is considered a 'project' is up to the laboratory performing the research. It could be the dating of a group of objects, but the laboratory can also decide to define a separate project for each object. Therefore, a project can have one or more objects associated with it. TRiDaS Object An object is the item to be investigated. Examples include: violin; excavation site; painting on a wooden panel; water well; church; carving; ship; forest. An object could also be more specific, for example: mast of a ship; roof of a church. Depending on the object type various descriptions are made possible. An object can have one or more elements and can also refer to another (sub) object. For instance a single file may contain three objects: an archaeological site object, within which there is a building object, within which there is a beam object. The list of possible object types is extensible and is thus flexible enough to incorporate the diversity of data required by the dendro community. Only information that is essential for dendrochronological research is recorded here. Other related data may be provided in the form of a link to an external database such as a museum catalogue. TRiDaS Element An element is a piece of wood originating from a single tree. Examples include: one plank of a water well; a single wooden panel in a painting; the left-hand back plate of a violin; one beam in a roof; a tree trunk preserved in the soil; a living tree. The element is a specific part of exactly one object or sub object. An object will often consist of more than one element, e.g., when dealing with the staves (elements) of a barrel (object). One or more samples can be taken from an element and an element may be dated using one or more derivedSeries. TRiDaS Sample A sample is a physical specimen or non-physical representation of an element. Examples include: core from a living tree; core from a rafter in a church roof; piece of charcoal from an archaeological trench; slice from a pile used in a pile foundation; wax imprint of the outer end of a plank; photo of a back plate of a string instrument. Note that a sample always exists and that it can either be physical (e.g. a core) or representative (e.g. a picture). A sample is taken from exactly one element and can be represented by one or more radii. TRiDaS Radius A radius is a line from pith to bark along which the measurements are taken. A radius is derived from exactly one sample. It can be measured more than once resulting in multiple measurementSeries. TRiDaS Measurement Series A measurementSeries is a series of direct, raw measurements along a radius. A single measurementSeries can be standardised or a collection of measurementSeries can be combined into a derivedSeries. The measurements themselves are stored separately as values. TRiDaS Derived Series A derivedSeries is a calculated series of values and is a minor modification of the 'vseries' concept proposed by Brewer et al (2010). Examples include: index; average of

20

a collection of measurementSeries such as a chronology. A derivedSeries is derived from one or more measurementSeries and has multiple values associated with it. TRiDaS Value A value is the result of a single ring measurement. Examples include: total ring width; earlywood width; latewood width. The values are related to a measurementSeries or a derivedSeries. In case of a measurementSeries the variable and its measurement unit (e.g. microns, 1/100th mm etc) are recorded as well.

21

Lihat lebih banyak...

Comentários

Copyright © 2017 DADOSPDF Inc.