GEOPHYSICALRESEARCHLETTERS,VOL. 27, NO. 18,PAGES2957-2960,SEPTEMBER15,2000
Swedish tree rings provide new evidence in support of a major widespread environmental disruption in 1628 BC H&kan Grudd Climate Impacts Research Centre, Kiruna, Sweden
Keith
R. Briffa
Climatic Research Unit, University of East Anglia, UK
BjSrn E. Gunnarson and Hans W. Linderholm Dept of Physical Geography, Stockholm University, Sweden
Abstract. Pine trees, recovered from a peat bog in southcentral Sweden, are used to develop a continuous, but "floating", 200-year tree-ring chronology. By wiggle matching
the Aegean Sea, which at the time had been approximately
•4C-datedto the 17th centuryBC. Subsequently, evidence
of a period of extremely low growth in Irish, English, and high-precision •4C determinations againstthe radiocarbon German oaks, starting in the year 1628 BC, was ascribed calibration curve, the chronologyis positioned at 1695 - 1496 to the sameevent [Baillie and Munro, 1988;Baillie, 1990]. BC with an uncertainty of 4-65 years. One major event, More recently,Kuniholmet al. [1996],haveidentifieda very denoted by four consecutive years with extremely narrow large growth anomaly in a tree-ring chronology from Anarings, indicative of highly unfavourable local tree-growth tolia (Turkey) that could alsobe dated to this time. Hence, conditions, occurred in the time window represented by the the current state of knowledge is of a major environmental chronology. This growth depressionis dated to 1637 BC disruption, possibly a major volcanic eruption in 1628 BC, (4-65) and may be tentatively ascribedto the samephe- that had widespread climatic effects in the mid-latitude regions of the Northern Hemisphere. nomenon that caused frost damage in trees in California
and a growth depressionin Europeanoak in 1628/27 BC,
Evidencefor the 1628/27 BC tree-ringeventin higherlat-
hence providing new evidence of a more northerly area of influence of this widespread phenomenon.
itudes has, until now, been lacking due to the few precisely dated tree-ring chronologiesacrossthis period. We now report new evidence indicating a strong effect in south-central Sweden apparently at this same date.
Introduction Tree rings are widely used as proxy records of environmental change. Different forms of growth stress can leave
Materials
and
methods
their imprint in the annual width of tree rings [Douglass, Hanvedsmossenis a peat bog situated 50 kilometres south 1920]. By combiningsamplemeasurementsfrom several of Stockholm in south-central Sweden. Large-scalepeat hartrees from within one site it is possible to construct a time series of environmental change that has annual resolution and, where it is continuousto the present day, absolute dat-
ing accuracy[Fritts, 1976]. Suchtree-ringchronologies can reachfar into the past [Spurket al., 1998] providingyearto-year information on local site conditions.
When, in 1984, LaMarche and Hirschboeck [1984]compiled a record of frost damaged tree-rings in bristlecone pine in the western USA, extending from the present back to 3435 BC, they found good agreement in the timing of implied frost events and volcanic eruptions known from historical sources. One very severefrost event in the prehistoric period dated to 1627 BC. Since this was the only frost-ring event in the second millennium BC it was tentatively attributed to
the cataclysmicBronzeAgeeruptionof Santorini(Theta) in •Also at Dept of PhysicalGeography,StockholmUniversity, Sweden
Copyright2000 by the AmericanGeophysicalUnion. Papernumber 1999GL010852. 0094-8276/00/1999GL010852505.00
2957
vesting at this site has exposed hundreds of pine trees in the contact zone between the peat and the substratum. Thus, the trees pre-date peat inception at this site. This ancient well-grown forest, radiocarbon dated to the 17th and 16th centuries BC, represents a 200-year period of relatively dry conditions when pine was able to germinate on a shallow dried-out fen-moor surface underlayed by fine-grained mineral soils. The forest was killed by a change to more humid conditions and was subsequently enveloped and preserved
by peat [Gunnarson,1999]. In this analysis we use 41 samples - full discs cut from
pine trunks. Tree-ringswere measuredto the nearest1/100 mm
in two
radii
in each disc.
The
two radii
were
cross-
checked for measurement error and averaged to produce a mean tree-ring growth seriesfor each tree. Mean tree-series were then crossdated following standard tree-ring methods
[Schweingruber, 1988] and givenrelative datesusingan arbitrary but consistent time scale. The aim of the present study is to highlight extreme events, i.e. abrupt short-term changes in the data. Thus, long-term influences, manifest as slowly evolving growth patterns, are of no interest, and can be filtered out leaving only the high frequency signal. This form of processing the data, known as "standardisa-
2958
GRUDD ET AL.: SWEDISH
TREE RINGS IN 1628 BC
Table 1. 14C-datesreferredto in Figure1
Cal. years BC -1750
Sample
Lab. ID
yr BP
la
Ring
Cal yr BC
a b c d e f
St14158 St13577 St13578 St14157 St14159 St14160
3435 3320 3200 3300 3360 3270
40 70 60 30 30 40
992 1007 1019 1060 1113 1168
1770-1630 1690-1500 1530-1390 1605-1505 1690-1530 1600-1450
-1700
,,i,,,,
-1650
-1600
-1550
i,,,,i,,,,i,,,,i
-1500
-1450
-1400
f,,,i,,,,i,,,,i
,
3500
3400
3300
3200
tion" [Fritts, 1976;Cookand Kairiukstis,1990],is routinely used in tree-ring chronology construction to highlight a specific environmental signal. Our choice of standardisation
3100
abc
methodinvolvedtwo steps: (1) transformationof the data
990
to logarithms in order to normalise the very large variabil-
mula:
=
-
where It is the index value for each year t; m and s are the mean and the standard deviation respectively. Z-scores outside
-93 are considered
as extreme
events in the series.
Six samples from the chronology, with precisely known
e
f
1030
1070
1110
1150
1190
Relative ring number in chronology
ity in the early juvenilephaseof tree growth; (2) filtering of each individual tree-ring series by fitting a flexible smoothing splinefunction with a 50% frequencycut-off of 100 years. This produces a set of indexed serieswhere most of the lowfrequency variation is filtered out. The indexed series were finally averaged to produce a mean chronology for the site. Extremes were identified and quantified by transforming the mean chronology to a series of z-scores, using the for-
d
I • , , I , , , I , q , I , , , I , i i j
Figure
1. Graphical representationof the wiggle match. See
also Table
1.
to our discussionbecause it bears directly on the extent to
whichwe canestablishwhetheror not the 1628/27 BC event, seen in California, central Europe and the British Isles, also had an influence on Scandinavian trees. The wiggle match is by no means perfect, since we excluded one outlier in the calculation. However, the radiocarbon calibration curve in
Figure2 is constructedfrom ten-yearblocksof wood[Stuiver and Becker,1993]whichmeansthat it can be regardedas a 10-yearlow passsmoothedrecord.Our •4C determinations
were made from 1-8 year blocks of wood and we should, therefore, expect our data to be more scattered than the The •4Cdeterminations werethenwigglematched[Pearson, ten-year smoothed calibration curve. Thus, we are confident that the 200-year chronology dates to 1695-1496 BC 1986],to the longradiocarboncalibrationcurve[Stuiverand with an uncertainty of 4-65 years.
separation(in years)and eachrepresentinga sampleof wood with 1-8 annual rings were radiocarbondated (Tab. 1).
Becker,1993] usingthe calibrationsoftwareOxCal [Bronk Ramsey,1994]. This operationis essentiallya least squares
The EuropeanOak tree-ringwidth chronologies [Baillie, 1990] and the bristlecone pine frost-ring record [LaMarche calculation of the combined errors, which significantly in1984],are both absolutelydated, i.e. there creases the datingaccuracy.Oneof the X4C-determinationsand Hirschbo.eck, is no uncertainty associated with the dates of individual (c in Tab. 1) was regardedas an outlier and was omitted from
the calculation.
Results
The wiggle match (Fig. 1) provideda date to within 4-65 calendaryearswith 95% probability (4-23 yearswith p=68%). Thus, the 200-year chronologycoversthe time periodfrom 1695- 1496 BC (4-65).
rings. Therefore, we can be certain that oak trees in different parts of central Europe and the British Isles were affected by a period of stressful growth conditions starting in the year 1628 BC. We can also be certain that pine trees at the tim-
berline
in the White
- 1650
1050-1053,have z-scoresbelow-3 (Fig. 2). The two most
were affected
- 1600
- 1550
- 1500
3
= o
anomalydates to 1637 BC 4-65 years (1625 BC 4-23 with p=68%). Minor low-growthanomaliescan be identifiedat relative rings 103d, 1086, 112d and 114{7 (Fig. 2).
-3
:;•:•:•:;•..t•:;::::•:•:•:•:;;•:•:;l•:;E?•:•:•:•:::•::•:•;•:•:;•? ß Y
. , . . 7000
, .... 7050
?700
, .... ?•50
•olativo ring no.
Discussion Our data from Hanvedsmossenrepresent an annually resolved and well-replicated record of environmental variation in a time window of 200 years in the second millennium BC. The dating precision of this "floating" chronologyis central
of California
Years BC (Wiggle matched)
We can identify one major anomalous growth event in the chronology: Four consecutive years, between relative rings extreme years have scores around-4, indicating a very extreme low-growth anomaly. The first year of this major
Mountains
by severe cold in the year 1627 BC. These lines of evidence
Figure 2. The mean chronologyhasbeentransformedto a time series of z-scores. The first year of the major growth reduction, starting at relative ring 1050, was wiggle matched to 1625 BC. The shaded area represents the 1• confidence interval for this date.
GRUDD
ET AL.: SWEDISH
point to a climate cooling in the mid-latitudes of the Northern Hemisphere starting in the year 1628 BC. Supportive, but less precise, tree-ring evidence of a climatic event at this time comesfrom a continuousbut "floating" 1,503-year tree ring chronology from Anatolia in Turkey, which has also been dated using the radiocarbon wiggle-match tech-
TREE
RINGS
IN 1628 BC
2959
the tree-ring signal is generally a secondary,climatic effect: Consequently, tree-ring derived attributions are only ever circumstantial.
The 1628/27 BC eventcausedan extremelyseveregrowth
depression in our data, lasting for about 4 years. Another less severedepression,but one that lasted 7 years, occurred nique [Kuniholmet al., 1996]. Similarly,the new 200-year one hundred years later, and several other shorter but nochronology from Hanvedsmossennow has a very unusual table growthreductionsare alsoapparent(Fig. 2), indicatgrowth reductionat 1637 BC +65 years (1625 BC +23). ing a high frequency of unfavourable tree growth periods. Our dating and the severe magnitude of this phenomenon This may be an indication of multiple volcanic eruptions suggestthat it can be ascribedto the 1628/27 BC event, in this 200-year time frame. Indeed, several other eruptions hence providing new evidence of a wider, more northerly beside Santorini have been approximately dated to this time area of influence.
The tree-ring evidence, thus, points to a major environ-
mental disruptionin 1628/27 BC having effect over much of the Northern Hemisphere. One likely candidate to have caused this effect is a major volcanic eruption. Large explosive volcanic eruptions are known to produce detectable cooling over a period of 1-3 years, even on temperatures
averagedacrossthe Northern Hemisphere[Angelland Korshover,1985]. Environmentallysensitivetree ringsfrom a wide range of geographical areas have been shown to display this temperature effect, caused by increased amounts of aerosolsin the stratospherewhich reduce solar radiation
[Vogelet al., 1990],and the ice coreshave acid spikesthat point to, as yet, unidentified volcanic activity. However, based on present knowledge, Santorini was the largest explosive volcanic eruption in this time. Conclusions A new, continuous but "floating", 200-year tree-ring chronology from Hanvedsmossenin south-central Sweden is radiocarbon dated, with the technique of wiggle matching, to the period 1695 - 1496 BC with an uncertainty of 4-65 years.
In the time window represented by the chronology, one major environmental event, probably a severe cooling, ocdata from Hanvedsmossenshow a very strong 4-year reduction in growth, presumably becauseof a severe cooling, con- curred at 1637 BC +65 years. This event is therefore consistent with the hypothesis of a major volcanic eruption in temporaneous with other tree-ring evidence of a severe climate anomaly at around 1628 BC, previously identified in 1628 BC. North America, western Europe and the Aegean, reinforcing Much of our present knowledgeon the timing and potenand extending the evidence to show an increased geographtial climate effectivenessof prehistoric volcanism comesfrom ical area of effect, now incorporating south-central Sweden. ice cores, through the measurements of acidity at different The evidence is consistentwith the hypothesis of a major depths which represent volcanic aerosol deposition through Northern Hemisphere volcanic eruption in 1628 BC, which time [Hammeret al., 1980;Zielinskiet al., 1994]. Ice-core may have been Santorini in the Aegean Sea. evidence of volcanic activity in the 17th century BC comes from three different deep coresfrom the Greenland Ice Sheet: Acknowledgments. This work was undertakenas part of The GRIP record has the largest acid peak over several hun- the ADVANCE-10K project funded by the EC, DGXII (ENV4dred years at an estimated date of 1636 BC, which correlates CT95-0127). We are also grateful to the Environment and Space
[Scuderi,1990; Joneset al., 1995;Briffa et al., 1998]. Our
to a peak at 1644 BC in the Dye 3 record [Clausenet al., 1997]. These dates have an uncertaintyof about 1-2 %,
ResearchInstitute in Kiruna, Sweden.
which means that they could derive from a single major volcanic eruption in 1628 BC. The GISP2 ice-core exhibits a
References
notable sulphatelayer at 1623 BC [Zielinski et al., 1994]. However, a significantly larger signal occurs at 1695 BC, which may be an indication of a larger uncertainty in the
GISP2 time scale[Clausenet al., 1997]. The 1628/27 BC tree-ring event and the ice core events have been assignedto the cataclysmic eruption of Santorini
(Theta) in the AegeanSea[LaMarcheandHirschboeck, 1984; Hammer et al., 1987; Baillie and Munro, 1988; Baillie, 1990; Zielinski et al., 1994; Kuniholm et al., 1996; Clausen et al.,
1997],basedon radiocarbondatesfrom the devastatedvillage of Akrotiri on the Island of Thera, which points to
a 17th centuryBC age for the eruption [Hubbertenet al., 1990]. However,the connectionbetweenSantoriniand the proxy evidence has been much discussed in the literature
[seeHardy and Renfrew, 1990; Baillie, 1996, 1998; Buckland et al., 1997; Manning, 1998; Zielinski and Germani,
1998a, 1998b, and referencestherein]. It is important to note that although ice cores directly record the deposition of acid aerosols,the attribution to specificvolcanoesmay be problematic, often only circumstantial. The linking of treering events to specific volcanoes is also problematic since
Angell, J.K., and J. Korshover, Surface temperature changes following the six major volcanic episodes between 1780 and 1980, J. Climatol. Appl. Meteorol., œ6, 937-951, 1985. Baillie, M.G.L., and M.A.R. Munro, Irish tree rings, Santorini and volcanic dust veils, Nature, 332, 344-346, 1988. Baillie, M.G.L., Irish tree rings and an event in 1628 BC, in Theta and the Aegean world III, edited by D.A. Hardy, and A.C. Renfrew, pp. 160-166, The Thera Foundation, London, 1990. Baillie, M.G.L., Extreme environmental events and the linking of the tree-ring and ice-core records, in J.S. Dean, D.M. Meko and T.W. Swetnam, eds., Tree Rings, Environment and Humanity: Proceedingsof the International Conference, Tucson, Arizona, 17-21 May 1994. Tucson Radiocarbon, 703-711, 1996. Baillie, M.G.L., Bronze Age myths expose archaeological shortcomings? a reply to Buckland et al. 1997, Antiquity, 72, 425427, 1998. Briffa, K.R., P.D. Jones, F.H. Schweingruber, and T.J. Osborn, Influence of volcanic eruptions on Northern Hemisphere summer temperature over the past 600 years, Nature, 393, 450-455, 1998.
Bronk Ramsey, C., Analysis of chronological information and radiocarbon calibration: The program OxCal, Archaeological Computing Newsletter, • 1, 11-16, 1994. Buckland, P.C., A.J. Dugmore, and K.J. Edwards, Bronze Age myths? Volcanic activity and human response in the Mediter-
2960
GRUDD ET AL.: SWEDISH
ranean and North Atlantic regions, Antiquity, 71, 581-593, 1997.
Clausen, H.B., C.U. Hammer, C.S. Hvidberg, D. Dahl-Jensen, and J. Steffensen, A comparison of the volcanic records over the past 4000 years from the Greenland Ice Core Project and Dye
3 Greenland ice cores, J. Geophys.Res., 102, (C12), 2670726723, 1997.
Cook, E.R., and L.A. Kairiukstis (Eds.), Methods of Dendrochronology: Applications in the Environmental Sciences, 394 pp., Kluwer Academic Publishers, Dordrecht, 1990. Douglass, A.E., Evidence of climatic effects in the annual rings of trees, Ecology, 1, 24-32, 1920. Fritts, H.C., Tree Rings and Climate, 567 pp., Academic Press, London, 1976. Gunnarson, B.E., A 200-year tree-ring chronology of pine from a raised bog in Sweden; Implication for climate change?, Ge-
ografiskaAnnaler, 81A, (3), 421-430, 1999. Hammer, C.U., H.B. Clausen, and W. Dansgaard, Greenland ice sheet evidence of post-glacial volcanism and its climatic impact, Nature, 288, 230-235, 1980. Hammer, C.U., H.B. Clausen, W.L. Friedrich, and H. Taubert, The Minoan eruption of Santorini in Greece dated to 1645 BC?, Nature, 328, 517-519, 1987.
Hardy, D.A., and A.C. Renfrew (Eds.), Theta and the Aegean World III, The Thera Foundation, London, 1990. Hubberten, H.-W., M. Bruns, M. Calamiotou, C. Apostolakis, S. Filippakis, and A. Grimanis, Radiocarbon dates from the Akrotiri excavations, in Thera and the Aegean World III, edited by D.A. Hardy, and A.C. Renfrew, The Thera Foundation, London, 1990. Jones, P.D., K.R. Briffa, and F.H. Schweingruber, Tree-ring evidence of the widespread effects of explosive volcanic eruptions,
Geophys.Res. Lett., 22 (11), 1333-1336,1995. Kuniholm, P.I., B. Kromer, S.W. Manning, M. Newton, C.E. Latini, and M.J. Bruce, Anatolian tree rings and the absolute chronologyof theeastern Mediterranean, 2220-718 BC, Nature, 381, 780-783, 1996.
LaMarche, V.C.J., and K.K. Hirschboeck, Frost rings in trees as records of major volcanic eruptions, Nature, 307, 121-126,
TREE RINGS IN 1628 BC Pearson, G.W., Precise calendrical dating of known growthperiod samplesusing a "curve fitting" technique, Radiocarbon, 28, 292-299, 1986.
Schweingruber,F.C., Tree Rings, Basics and applications of Dendrochronology,Reidel, Dordrecht, 1988. Scuderi, L.C., Tree-ring evidence for climatically effective volcanic eruptions, Quaternary Research, $4, 67-85, 1990. Spurk, M., M. Friedrich, J. Holmann, S. Remmele, B. Frenzel, H.H. Leushner, and B. Kromer, Revisions and extension of the Hohenheim Oak and Pine chronologies: New evidence about
the timing of the Younger Dryas / Preboreal transition, Radiocarbon,40, (3), 1998. Stuiver, M., and B. Becker, High-precision decadal calibration of the radiocarbon time scale, AD 1950-6000 BC, Radiocarbon,
35, (1), 35-65, 1993. Vogel, J.S., W. Cornell, D.E. Nelson and J.R. Southon, Vesu-
vius/Avellino, one possiblesourceof seventeenthcentury BC climatic disturbances, Nature, 344, 534-537, 1990. Zielinski, G.A., P.A. Mayewski, L.D. Meeker, S. Withlow, M.S. Twickler, M. Morrison, D.A. Mese, A.J. Gow, and R.B. Alley, Record
of volcanism
since 7000 B.C.
from the GISP2
langesthe 1620sBC age for the Santorini (Minoan) Eruption, J. of ArchaeologicalScience, 25, 279-289, 1998a. Zielinski, G.A., and M.S. Germani, Reply to: Correction. New GISP2 Ice-core evidence supports 17th century BC date for
the Santorini (Minoan) eruption, J. of Archaeological Science, 25, 1043-1045, 1998b.
H. Grudd, Climate Impacts Research Centre, BjSrkplan 6A, S-981 42 Kiruna, Sweden, e-mail:
[email protected] K.R. Briffa, Climatic Research Unit, School of Environmental Sciences,University of East Anglia, Norwich NR4 7TJ, UK B.E. Gunnarson and H.W. Linderholm, Dept of Physical Geography, Stockholm University, S-10691 Stockholm, Sweden
1984.
Manning, S.W., Correction. New GISP2 Ice-core evidence sup-
ports 17th century BC date for the Santorini (Minoan) eruption: Responseto Zielinski & Germani (1998), J. of Archaeological Science, 25, 1039-1042, 1998.
Green-
land ice core and implications for the volcano-climate system, Science, 264, 948-952, 1994. Zielinski, G.A., and M.S. Germani, New ice-core evidence chal-
(ReceivedJune 4, 1999; revised August 27, 1999; acceptedFebruary 7, 2000.)
