SNorwQ: Difference between revisions

From Cybis Wiki
Jump to navigation Jump to search
(→‎The data retrieving process: using Log-labels instead will eliminate the need of spreadsheet)
Line 175: Line 175:


===The data retrieving process===
===The data retrieving process===
In some reports the data is in a [[.fh]]-like (or [[Tucson-format|Tucson]]-like format, which needs some manual editing to become readable by [[CDendro]]. A majority of the reports contain data as curve diagrams in various resolutions, in many cases excellent, in other rather bad. To extract good images from the pdf-files GSview was used.<ref>See {{enWP|GSview}}. The actual pages were saved as .jpg (gray) images at 600 dpi</ref> Data from such curve diagrams could rather easily and usually with really good precision be retrieved using [[CooRecorder]]. As some of the reports contain the same data as tables and as diagrams, it has been possible to figure out the way the data was transformed for the curve writing process. The formula which was found to have been used is:
In some reports the data is in a [[.fh]]-like (or [[Tucson-format|Tucson]]-like format) which needs some manual editing to become readable by [[CDendro]]. A majority of the reports contain data as curve diagrams in various resolutions, in many cases excellent, in other rather bad. To extract good images from the pdf-files GSview was used.<ref>See {{enWP|GSview}}. The actual pages were saved as .jpg (gray) images at 600 dpi</ref> Data from such curve diagrams could rather easily and usually with really good precision be retrieved using [[CooRecorder]], see [[Danish National Museum]], for detalis. The method used for retrieving the data from those diagrams having just "1" as y-axis label, was afterwards found to be a bit more complicated than necessary. Just using the CooRecorder calibration label "'''SCALE 0.535 1 Log'''" would have given the same result. The optimal label, however,  for those curves are "'''SCALE 0.52 1 Log'''". The difference between the used label and the optimal one leads to a small underestimation of the widest rings and a small overestimation of the narrowest rings, -6% for ring of 4 mm width and +5.5% for rings at 0.3 mm. Rings of 1 mm will be correctly estimated). This divergence will be of very small, if any, relevance as far as the collection is used for cross dating purpose only. 


<math>D=1+k \ln(W) </math> (W = Ring width, D = diagram value, k = a factor found to be approximately 1.6)
The diagrams of a few reports has two y-axis labels given in 1/100 mm, e.g 20 and 200 (i.e 0.20 mm and 2.00 mm), are not affected by this problem.
 
Ring width values (W-valus) could consequently be restored from the retrieved data (D-values) by the corresponding formula:
 
<math>W=e^{D-1\over k}</math>
 
This conversion was done in a Spreadsheet (OpenOffice.org Calc)<ref>{{enWP|OpenOffice.org_Calc}}. (Microsoft Excel would work as well.)</ref>
 
The calibration label useful for most of the diagrams is: '''SCALE 0 1 Prop'''. The double point should be at the mark labeled "1" and the unlabeled mark below. Note that this will generate negative values for narrow (<0.535 mm) rings, so the "raw" wid-files needs the conversion described above before they are used with CDendro.
 
The diagrams of a few reports are labeled in a different way. To avoid the confusing appearance of what looks as negative ring width values, two levels of corresponding true ring width levels are labeled (in 1/100 mm). But as the diagram still shows D-values according to the formula above, the given label values have to be converted to D-values before they are useful as calibration labels in CooRecorder. E.g in the report [http://www.nnu.dk/dendro/A8827%20Rap%20Oddernes%20kirke.pdf A8827] the labels are 20 and 200 (i.e 0.2 mm and 2.0 mm). W=0.2 gives D=-1.575 and W=2.0 gives D=2.109. The CooRecorder label for this diagram therefore should be: '''SCALE -1.575 2.109 Prop'''. In other reports other label values are given, which means that proper values for the CooRecorder labels have to be estimated from case to case.
 
In many reports the average value of all measurements of an actual series is given. Normally this value is also found in the commentary to the samples in SNorwQ. If the average value of the retrieved data is also estimated it gives an idea of the accuracy of the conversion method. Often this value is inserted in the comment.
 
'''Note:''' This conversion was later realized to be equivalent to the use of the label '''SCALE 0.535 1 Log''', which eliminates the need of any spreadsheet operation.


==Notes==
==Notes==

Revision as of 23:28, 19 February 2010

SNorwQ is an oak reference for southern Norway AD 1333-2006. It consists of selected samples from two internet sources. Most of the recent trees (about 70) are from ITRDB (.cat-files) ftp://ftp.ncdc.noaa.gov/pub/data/paleo/treering/updates/contributions/christensen_denmark_oak/norge/ By Kjeld Christensen and Kent Havemann. The files used are those from the southernmost fylken (counties) of Norway, Aust Agder and Vest Agder. Some recent samples and all the historical are from several dendrochronological dating reports mainly by Claudia Baittinger and Orla Hylleberg Eriksen at the Danish National Museum i Copenhagen (in cooperation with Norwegian antiquarian authorities in Vest Agder Fylke), available at http://www.nnu.dk/dendro/dendrorapsdk.htm (and http://www.vaf.no/hoved.aspx?m=1331&amid=6707 )(PDF-files i Danish, see detailed list below). All samples selected for SNorwQ have been individually scrutinized before they were added to the collection.

The main reason for collecting all those samples, was to create a south Norwegian master chronology. Especially in this perspective we really appreciate the great generosity of the Danish National Museum, who do share their data in a way which invites us as outsiders to participate and to construct our own references based on their data. Although our SNorwQ reference surely is a lot weaker (and shorter?) than the reference the Museum has built, it will be of great value in helping us recognize Norwegian wood, e.g. in the former Danish/Norwegian provinces in southern and western Sweden, or just to get an idea about "the dendrochronological distance" to this area during different time spans. In the reports as well as in the ITRDB it is stated that the material may be used freely as long as the publications are referred to, and the authors are acknowledged. This seems to be compatible with the CC-BY-SA license used by this wiki.

Versions

Recent samples from the ITRDB collection

Recent and historical samples from NNU reports

List of samples from NNU-reports used. In some reports the ring width measurment data are given in numerical form, in others in form of curve diagram which can be retrieved but not without some data degeneration. In some the same data is given in both forms. In those cases the numerical data is used. Note that some samples has been truncated.

  • NNU rapport nr. 1 ● 2006[1] A8457 (or A8607), Kvelland stabbur (i.e granary built on wooden pillars), by Claudia Baittinger. Data from table. 58°9′N 7°6′E
    • N0640019 1597-1663
    • N0640029 1627-1682
    • N0640039 1617-1676
    • N0640049 1573-1674
    • N0640059 1571-1677
    • N0640069 1611-1680
    • N0640079 1548-1657
  • NNU rapport nr. 3 ● 2006[2], A8614 Greipsland stolpehus (House on pillars), by Claudia Baittinger and Thomas Bartholin. Data from tables. 58°3′N 7°3′E
    • N0670029 1772-1822
    • N0670039 1766-1823
    • N0670049 1765-1822
    • N0670069 1761-1823
    • N0670079 1758-1824
    • N0670109 1755-1823
  • NNU rapport nr. 6 ● 2006[3] A8611 Stabbur Eiken Tunet in Hægebostad. by Claudia Baittinger and Thomas Bartholin. Data from table. 58°29′N 07°13′E
    • n0651019 1659-1812 A1 stabbe nordvæg (truncated in both ends, Block 28-180)
    • n0651029 1800-1862 A2 stabbe nordvæg
    • n0651039 1707-1862 A3 stabbe nordvæg
    • n0651049 1657-1863 A4 stabbe sydvæg
    • n0651059 1797-1862 A5 stabbe sydvæg (Truncated: Oldest part depressed)
  • NNU rapport nr. 8 ● 2006[5], A8627, fartøjsrester fra (Ship remains from) Lindebøkilen, Kristiansand kommune, Vest-Agder fylke
    • 02300029 Not included according to low corr (and a little bit better corr towards HallandQ than towards SNorwQ!)
  • NNU rapportblad 2006, CB ● 1[6], A8613. Living trees. By Claudia Baittinger. (Data retrieved from low resolution curves). 58°3′N 7°3′E
    • N0430079 1820-1984 From the sawmill Møllsag A/S ved Holum. The tree was said to be harvested in an area east of Kristiansand.
    • N0660019 1828-2005 Våjeskauen, Kristiansand kommune - Sample from living tree
  • NNU rapport nr. 1 ● 2007[8] A8611 Fjøs (cowshed), by Claudia Baittinger and Thomas Bartholin. Retrieved curve data. 58°29′N 07°13′E
    • n0651t01 1659-1746 Mean of several samples. Truncated. Only Block 117-203 (0 is youngest ring) is included as the rest partly seems to be severely disturbed.
    • n0651t02 1781-1862 Mean of several samples
    • N0652m02 1761-1869 Mean of several samples
  • NNU rapport nr. 10 ● 2007[9] A8611 (or A8672) Grindheim kvernhus (House for small water mill), by Claudia Baittinger. Retrieved curve data. 58°26′N 7°26′E
    • N0730019 1557-1715 (01) [Block 33-190 of original included. The youngest 33 years too low corr]
    • N0730039 1565-1763 (03)
    • N0730059 1592-1730 (05)
  • NNU rapport nr. 12 ● 2007[10], A8708 Møskeland stabbur, by Claudia Baittinger. Data retrieved from curve diagrams. 58°9′N 7°4′E
    • N0800029 1773-1821 2. stok over gulv 46+67cm fra hjørne mod n-ø {2 pr.}
    • N0800079 1744-1825 3. stok over dør 150 cm fra hjørne mod nord-vest
  • NNU rapport nr. 5 ● 2008, A8611 (or 7285) Søgne gamle kirke (Søgne old church). Link is broken on the NNU-page, but available here[11] (Vest-Agder Fylke). By Claudia Baittinger. Data from table. 58°6′N 07°51′E
    • N0890019 1515-1599 (Block 21-104 - missed ring at about AD1600) Bindingsverk i skipets østvegg. 1.andreaskors i. øverste fag, fra nord. Ene leddet.
    • N0890029 1534-1626 Bindingsverk i skipets østvegg. 2. andreaskors fra nord, øverste fag.
    • N0890039 1544-1618 radii y+b, Bindingsverk i skipets østvegg. 2. andreaskors fra nord, øverste fag. Andre leddet.
    • N0890049 1507-1622 Bindingsverk i skipets østvegg. 3. andreaskors fra nord; øverste fag.
    • N0890059 1491-1557 radii b+c [Truncated in both ends]; Bindingsverk i skipets østvegg. Stolpe nr. 2 fra nord-østre hjørne.
    • N0890069 1423-1584 Bindingsverk i skipets østvegg. Remmen.
    • N0890079 1519-1620 [Same stem as 3] Bindingsverk i skipets østvegg. Andreaskors nr. 1 i øverste fag, fra nord. Andre leddet.
  • NNU rapport nr. 23 ● 2008[13] A7061 Bjorvatn Stabbur (i.e granary built on wooden pillars), by Claudia Baittinger. Data retrieved from curve diagrams. 58°31′N 08°28′E
    • N0271019 1566-1716 (B1 tverrbjelke)
    • N0271029 1493-1716 (B2 tverrbjelke)
    • N0271t03 1538-1696 Mean of several. (The identity is from the the retrieved curve but missing in the 'Katalog'-section of the file. Maybe confused with N0271t02 in the catalog but without curve?)
  • NNU rapport nr. 9 ● 2008[17] (A version including data avalable at www.vaf.no[18]!), A8740 Bjerland stabbur 1-4 and fjøs/driftsbygning (cowshed/economy building), by Claudia Baittinger. Data from table. 58°23′N 7°29′E
    • N0850019 1642-1743 stabbur 1
    • N0850029 1458-1606 stabbur 2
    • N0850039 1637-1735 stabbur 3 radii c+d
    • N0850049 1615-1721 stabbur 4
    • N0851019 1774-1869 fjøs/driftsbygning 1
    • N0851029 1854-1930 fjøs/driftsbygning 2
    • N0851039 1879-1930 fjøs/driftsbygning 3
    • N0851069 1883-1930 fjøs/driftsbygning
  • NNU rapport nr. 12 ● 2008[21], A8743 Lundevoll (Kvås) kvernhus (mill), by Claudia Baittinger. Data retrieved from curve diagrams. 58°13′N 7°11′E
    • N0880019 1737-1881
    • N0880029 1755-1851
    • N0880039 1744-1844
    • N0880049 1762-1877
    • N0880059 1839-1882 [2 missing rings caused by a covering label in the curve diagram]
    • N0880069 1814-1881
    • N0880089 1810-1872
    • N0880099 1805-1881
    • N0880109 1818-1881
  • NNU rapport nr. 17 ● 2008[24], A8717 Tjersland, by Claudia Baittinger. Data retrieved from curve diagrams. 58°9′59″N 7°3′17″E (original place)
    • N0820019 1777-1839 C03 - 1. stokk over sleppverk; 5 cm fra nordre ende av stokken
    • N0820109 1796-1831 C12 - Stokk nr. 5 fra gulv; 124 cm fra hjørne mot øst
  • NNU rapport nr. 11 ● 2009[26], A8827 Oddernes kirke (Oddernes church); bjælke 1-4 (Beam 1-4), by Claudia Baittinger. Data retrieved from curve diagrams.
    • N1020019 1723-1806 bjælke 1
    • N1020029 1741-1825 bjælke 2
    • N1020039 1729-1823 bjælke 3
    • N1020049 1675-1824 bjælke 4
  • NNU rapportblad 2009 OHE ● 20[29], A8842 Holt kirke, kiste (chest), by Orla Hylleberg Eriksen. Data retrieved from low resolution curve diagram - degeneration of data expected.
    • N1110019 1532-1663 [TRUNCATED: A probably 1 year miss-dated block is removed]
  • NNU rapportblad 2009 OHE ● 27[30], A8854 svill fra stabbur fra (sill from a Granary on pillars from) Frikstad, by Orla Hylleberg Eriksen. Data retrieved from low resolution curve diagram (some degeneration of data expected).
    • N1150109 1374-1473 mean of 2 samples from the same stem [TRUNCATED: Block 20-118 included]

The data retrieving process

In some reports the data is in a .fh-like (or Tucson-like format) which needs some manual editing to become readable by CDendro. A majority of the reports contain data as curve diagrams in various resolutions, in many cases excellent, in other rather bad. To extract good images from the pdf-files GSview was used.[31] Data from such curve diagrams could rather easily and usually with really good precision be retrieved using CooRecorder, see Danish National Museum, for detalis. The method used for retrieving the data from those diagrams having just "1" as y-axis label, was afterwards found to be a bit more complicated than necessary. Just using the CooRecorder calibration label "SCALE 0.535 1 Log" would have given the same result. The optimal label, however, for those curves are "SCALE 0.52 1 Log". The difference between the used label and the optimal one leads to a small underestimation of the widest rings and a small overestimation of the narrowest rings, -6% for ring of 4 mm width and +5.5% for rings at 0.3 mm. Rings of 1 mm will be correctly estimated). This divergence will be of very small, if any, relevance as far as the collection is used for cross dating purpose only.

The diagrams of a few reports has two y-axis labels given in 1/100 mm, e.g 20 and 200 (i.e 0.20 mm and 2.00 mm), are not affected by this problem.

Notes

  1. ↑ WebCite archive
  2. ↑ WebCite archive
  3. ↑ WebCite archive
  4. ↑ WebCite archive
  5. ↑ WebCite archive
  6. ↑ WebCite archive
  7. ↑ WebCite archive
  8. ↑ WebCite archive
  9. ↑ WebCite archive
  10. ↑ WebCite archive
  11. ↑ WebCite archive
  12. ↑ WebCite archive
  13. ↑ WebCite archive
  14. ↑ WebCite archive
  15. ↑ WebCite archive
  16. ↑ WebCite archive
  17. ↑ WebCite archive
  18. ↑ WebCite archive
  19. ↑ WebCite archive
  20. ↑ WebCite archive
  21. ↑ WebCite archive
  22. ↑ WebCite archive
  23. ↑ WebCite archive
  24. ↑ WebCite archive
  25. ↑ WebCite archive
  26. ↑ WebCite archive
  27. ↑ WebCite archive
  28. ↑ WebCite archive
  29. ↑ WebCite archive
  30. ↑ WebCite archive
  31. ↑ See Wikipedia (English) article about GSview. The actual pages were saved as .jpg (gray) images at 600 dpi
Retrieved from "http://cdendro.se/wiki/index.php?title=SNorwQ&oldid=1813"