Correlation between harvest results 1683-1780 and tree ring data
Dendrochronological study
by Torbjörn Axelson, August 2011
Svensk version


Occasionally I found a table containing a classification for each year during the period 1683-1780 according to harvests results in Sweden, done 1783 by someone named Ekman.1 He is using seven classes, which I will replace with numbers in order to allow comparison with dendrodata:
After this replacement I have a "dendro-style" measurement series, for which I did an ordinary test towards all available pine ring width chronologies from Scandinavia,2 using the normalization method "proportion of last two years", i.e only the short term variation (year to year) is compared. The result of this exercise was a bit amazing: Surprisingly strong correlation values occurred towards especially a chronology from western Gästrikland3 (T-värde > 6), but rather good also towards chronologies for Dalarna and southeastern Småland. The first about 15 years, however, gives significant weaker values than the later period. This disagreement seems to be caused by the thinnest ring, the one for 1688, occurs one year after the reported crop failure 1687 - a delayed response in tree rings seems likely. Also the very last years, shows a bit lower correlation again.

It is unclear to me why the correlation towards exactly Gästrikland among all areas in Sweden, gives such a strong correlation towards Ekmans classification. It may have any of various reasons. Gästrikland may at this time have been a more prominent producer of cereals when today, according to the fertilization possible due to a strong cattle stock. So the situation in this and similar areas may have had some stronger impact on the general judgment on harvests, than we may expect? It is also possible that Ekman, about whom I do not know anything, may have been biased in favor for the situation in this particular area - maybe he lived somewhere around or he used sources from there? In any case this pine-chronology from 200 meter above sea level, seems to record the balance between temperature and precipitation in a way strongly affecting also the needs of the cereals during at least great parts of 18'th century.

A diagram showing the correlation between the pinus curve from Kungsberg in Gästrikland (green and red) and Ekmans harvest classification (blue and black).
Red and black curves are normalized. The years of crop failure are marked.



Cybis CDendro, Algorithm: Proportion of last two years growth (2,0,T
Correlations between references and
SKRDekm dated to 1780 with corr >= 0.25 and with overlap >= 90
Results sorted according to decreasing correlation coefficient values.
                    T-  Over
           Corr   Test   lap
Kungsber   0.53   6.07    97   1458-2005  Kungsberg-Järbo-Lumsheden, Gästrikland                  PISY
Bingsjo_   0.46   5.09    97   1400-1859  Bingsjö Dalarna, Klockargårdens storstuga               PISY
SE003-Hö   0.45   4.84    92   1688-1998  Högelycke, Nybro (Extraterastrial SE003)                PISY
SödraDal   0.44   4.84    97   1362-2005  Södra Dalarna                                           PISY
se007      0.44   4.75    97    931-1888  Dalarna region (Lunds Universitet, "extraterastrial")   PISY
swed305    0.43   4.60    97   1451-2002  Björbo, Dalarna                                         PISY
BratPISY   0.42   4.48    97   1153-1904  Mälardalen area, publ. Bråthen                          PISY
ESVEALAN   0.39   4.09    97   1661-2004  East Svealand Aneby+Askersu+Bettna+Sissham+Saltsjöba    PISY
HallandP   0.34   3.56    97   1605-2006  Halland                                                 PISY
swed022w   0.33   3.35    97   1128-1987  Gotland (Schweingruber)                                 PISY
Petmyra    0.27   2.63    90   1690-2002  Petmyra (bogg), Björbo, Dalarna                         PISY
Savolinn   0.26   2.64    97    535-2002  Savonlinna-området. Sammanslagna från  ITRDB            PISY
swed302    0.25   2.56    97   1583-1995  NAMDO  1 Naemdoe Stockholm archipelago (Updated 2/2003) PISY

SKRDE  1 Ekmans skördeomdöme 1683-1780                       
SKRDE  2 Sweden                                                    1683 1780
SKRDE  3 Ekman 1783 (interpretation: T. Axelson)
SKRDE   #### I have interpreted Ekmans judgments as: missväxt (failure) 50; dyr (dearth) 60; ringa (small) 65; knapp (little) 70; svag (weak) 75; medelmåttlig (normal) 80; ymnig (rich) 100
SKRDekm 1683    80    65    80    70    50    70    80
SKRDekm 1690   100   100    80    70    60    70    70    50    50    60
SKRDekm 1700   100   100    80    80   100    80    80    80    80    50
SKRDekm 1710    80    80   100   100    80   100    80    70    70    50
SKRDekm 1720    80    80   100    80    70    80    50    70    80   100
SKRDekm 1730    80    80   100    80    80    80    80   100   100    80
SKRDekm 1740    70    50    70    80   100    65    70    50    75    60
SKRDekm 1750   100    80    80   100    80    80    70    75    80   100
SKRDekm 1760    80    70    50    75    70    80    80    80   100   100
SKRDekm 1770    80    50    70    80    80    70    80   100    80   100
SKRDekm 1780    70   999

Notes

1. Ekmans skördeomdömen 1683-1780 i Sverige, posted by Stefan Lundgren 04 jul 2005, 13:25 at http://forum.skalman.nu/viewtopic.php?f=44&t=19914&p=292484, with reference to Per-Erik Brolin in Statistisk tidskrift (Statistical review) nr 10, 1954
2. Also oak and spruce chronologies were tested but no strong correlations were found.
3. Jan-Olov Språng: Kungsberg PISY, Cybis wiki, (Pinus sylvestris ring width collection from the area around Järbo in western Gästrikland, about 200 m a s l, 60º46´N, 16º28´E.) Master compiled by CDendro NegExp-detrend function

This and other dendrochronological studies by T. Axelson are found at taxelson.se/dendro/obj/